00:01:07 <fizzie> Oh, wow, that was unexpected: I just copied the old triple-head xorg.conf in place, and now it works. Some later xorg (or kernel) update must've fixed whatever was wrong with it.
00:11:31 <fizzie> Don't worry, the window manager (Awesome) is broken anyway. (They use a Lua script as the config file, and change pretty much all interfaces for every minor revision; I just upgraded from 3.3 to 3.4 and everything's broken.)
00:18:13 <fizzie> Well, it's a tiling wm, with the whole "tags" thing. I could even like it if they'd selected something else than Lua for their scripting lanugage.
00:19:03 -!- ehird has joined.
00:20:31 -!- Rugxulo has joined.
00:21:46 -!- kar8nga has quit (Remote closed the connection).
00:22:04 <Sgeo> Is Pyglet any good?
00:22:40 <ehird> 12:36:13 * AnMaster wonders how sane it is to burn an iso file when the iso file is on an nfs mount
00:22:47 <ehird> 100 Mb internet is faster than a CD.
00:23:05 <ehird> Sgeo: it is, I recall, unmaintained.
00:23:09 <Rugxulo> ehird, dumb question but what does "roger the googly" mean??
00:23:14 <Rugxulo> (say it mentioned on Family Guy)
00:23:34 <ehird> Rugxulo: I don't know? Family Guy is retarded
00:23:54 <Rugxulo> and MS won't sponsor it for Win7 anymore (awww, too bad)
00:24:04 <fizzie> AnMaster: I did think about that; it was disrecommended by someone, though.
00:24:31 <Sgeo> o.O Why is 0, 0 the LOWER left corner? In graphics stuff, isn't it typically the upper left?
00:24:34 <fizzie> I don't remember at all, just that someone voiced a negative opinion.
00:24:59 <Sgeo> http://www.pyglet.org/doc/programming_guide/image_viewer.html "The arguments (0, 0) tell pyglet to draw the image at pixel coordinates 0, 0 in the window (the lower-left corner)."
00:25:05 <AnMaster> Sgeo, and in math you generally have positive y up
00:25:08 <fizzie> Lower-left isn't so terribly uncommon either.
00:25:25 <fizzie> It does feel a bit more "mathy", yes.
00:25:26 <Rugxulo> didn't OS/2 use something odd like the top right??
00:25:32 <ehird> fizzie: dwm is cool innit
00:25:34 <AnMaster> fizzie, I think we should use polar coordinates
00:25:43 <ehird> TGA uses 0,0 = bottom right
00:26:07 <ehird> you can set a field
00:26:08 <AnMaster> still, I want an image format with polar coordinates
00:26:10 <ehird> to flip the coordinates
00:26:38 <AnMaster> ehird, Hm what is tga meant for? Maybe for "print this while it is being sent"?
00:26:44 <AnMaster> thus maybe making sense or such
00:26:55 <ehird> TGA is just for everything.
00:26:56 <ehird> Assuming you don't need compression.
00:27:00 <ehird> brb, trying to switch to wifi
00:30:16 * Rugxulo is surprised ehird doesn't like Family Guy, seems right up his alley
00:36:12 -!- ehird has quit (Read error: 145 (Connection timed out)).
00:36:15 -!- Rugxulo has quit ("ERC Version 5.3 (IRC client for Emacs)").
00:48:41 -!- ehird has joined.
00:48:52 <ehird> 13:38:56 <AnMaster> wow he uses ion? XD
00:48:52 <ehird> There's nothing wrong with ion, the author just has a bad personality
00:48:52 <ehird> 14:13:53 <fizzie> Isn't it against Ion's philosophy to do multihead, though? I mean, the reason given for not supporting Xinerama is "I'm not going to waste my time rubbing your unecological penis enlargement"; admittedly it supports plain X multihead, but I don't see how that would make it any more ecological.
00:48:56 <ehird> He's said on the mailinglists just to use regular multihead, i.e. he's not going to expend *effort* making it work for multiple monitor users
00:48:59 <ehird> 15:50:29 <AnMaster> fork ion?
00:49:01 <ehird> 15:51:40 <AnMaster> someone should. It would annoy the hell out of the author
00:49:05 <ehird> 15:57:43 <AnMaster> bsmntbombdood_, switch WM seriously.
00:49:09 <ehird> 15:58:21 <AnMaster> fizzie, ion wasn't the first tiling one was it?
00:49:11 <ehird> No; pwm, ion's successor, was the first tabbing WM
00:49:13 <ehird> 15:59:20 <AnMaster> dwm or such I would guess
00:49:17 <ehird> (There are no real alternatives to ion if you want tabbing+tiling)
00:49:19 <ehird> Also, stop talking about ion
00:49:21 <ehird> Stupid being disconnected while logreading
00:49:23 <ehird> Also stupid #esoteric being stuck on boring topics for ages
00:49:27 <ehird> 1.48 Mb/s download, 87 ms ping. Wi-Fi to router/modem downstairs, with something like 3 Mb/s theoretical maximum.
00:49:30 -!- FireFly has joined.
01:04:58 -!- Gracenotes has joined.
01:09:29 <ehird> "Sorry, the GeoCities web site you were trying to reach is no longer available."
01:09:42 <ehird> RIP GeoCities 1994—2009
01:10:03 <ehird> Or should I say Beverly Hills Internet?
01:13:13 <ehird> I wonder how to mount a loopback ext2 on OS X.
01:16:20 <Sgeo> Yesterday, I felt nostalgia for a website. It wasn't Geocities
01:16:59 <Sgeo> It's just, I'm usually easily affected by nostalgia, and you'd think that I'd feel nostalgia for Geocities
01:17:40 <ais523> I feel nostalgic for GeoCities
01:17:57 <ais523> I did a Google search today, and the fourth result, and the first useful one for me, was actually a Geocities website
01:18:24 <ais523> looks like I found it just in time
01:18:36 <ehird> I'd say boycott Yahoo!, but nobody uses Yahoo!.
01:18:37 <Sgeo> What was it, if I may ask?
01:18:40 <ehird> Apart from you, for emergency email.
01:18:44 <ais523> Sgeo: a massive list of prime numbers
01:18:48 <ehird> (Seriously; nobody uses Yahoo!!)
01:18:49 <ais523> ehird: also, I use Yahoo!
01:18:54 <ehird> (Also, FUCK THEIR NAME)
01:19:00 <ehird> ais523: why? it gives crappy resultsst
01:19:04 <ais523> for a non-work mail account
01:19:11 <ehird> [00:18] ehird: Apart from you, for emergency email.
01:19:14 <ehird> the agora one, yes?
01:19:16 * Sgeo likes girls. Not names
01:19:23 <ehird> that's a pretty minnor use
01:19:29 <ais523> because the SMTP server on bham.ac.uk keeps going down
01:19:37 <ehird> "Sgeo likes girls. Not names"
01:19:43 <ais523> making it rather hard to send emails, and in nomic timing is often important
01:19:45 <ehird> it doesn't mean a thiing
01:19:46 <Sgeo> <ehird> (Also, FUCK THEIR NAME)
01:19:57 <ehird> Sgeo: oh. now if only it was funny!
01:21:11 -!- cmeme has quit (Read error: 110 (Connection timed out)).
01:23:59 -!- bsmntbombdood has joined.
01:28:27 <ehird> "Simple generic tabbed fronted to xembed aware applications, originally designed for surf but also usable with many other application, i.e. uzbl, urxvt and xterm"
01:28:30 <ehird> I love you, suckless
01:29:33 <ehird> 841 lines; quite big for suckless
01:29:46 <ehird> tabbed \- simple webkit-based browser
01:35:30 -!- coppro has joined.
01:35:43 <ehird> reporting a bug to suckless was refreshingly simple.
01:36:06 <ehird> send to dev+subscribe-nomail@suckless.org, reply to its response. email dev@suckless.org
01:37:19 <ehird> tabbed \- simple webkit-based browser
01:37:23 <ehird> obviously copied from the surf manpage
01:37:47 <ehird> (that quote from the manpage was my entire message body :P)
01:38:10 -!- oerjan has quit ("leaving").
01:40:09 <ehird> hmm, t'isn't working; wonder what i broke
01:40:12 <pikhq> Hmm. What use could I get out of tabbed, anyways?
01:40:27 <ehird> pikhq: using a non-tabbed browser; tabs are a window management function
01:40:28 <pikhq> (I specifically, not "what does it do")
01:40:42 <ehird> e.g. http://surf.suckless.org/
01:40:52 <ehird> "It is able to display websites and follow links."
01:41:04 <pikhq> ehird: Tempting. I am always on the lookout for less-sucky web browsers.
01:41:21 <ehird> surf is adding one of those mouseless-link-clicky things sometime I belieeve, if that's yoru thing
01:41:34 <ehird> http://man.suckless.org/surf/1/surf pretty much lists everything it can do
01:41:41 <pikhq> That's one of the things I do with Conkeror, yes.
01:41:44 <ehird> admittedly, webkit/gtk handles basically the entire page
01:42:24 <ehird> http://surf.suckless.org/patches/history history is an optional patch :-D
01:42:32 <ehird> cat ~/.surf/history | sort -r | uniq | dmenu -l 10 -b -i | xprop -id `cat ~/.surf/id` -f _SURF_URI 8s -set _SURF_URI
01:42:35 <ehird> nice history viewer
01:42:49 <pikhq> It appears to be sufficiently non-suck that I may give it a shot in the near future.
01:42:49 <ehird> would be better with dmenu-vertical, though
01:43:25 <ehird> I love the suckless guys
01:44:24 <pikhq> Hmm. Wmii is using 9p?
01:44:31 <pikhq> That's kinda sweet.
01:44:40 <ehird> The suckless guys are huge plan9 fans.
01:44:44 <ehird> wmii depends on plan9port, iirc
01:44:46 <ehird> or, wait, just the 9base subset
01:45:10 <ehird> werc, their website tool, is written entirely in rc (plan9 shell)
01:45:29 <ehird> pikhq: dwm is a bit more suckless than wmii, though, i'd say
01:45:53 <ehird> of course there can be a thing as too much simplicity, but as you use ratpoison i somehow i doubt that applies
01:46:08 -!- puzzlet has quit (Read error: 104 (Connection reset by peer)).
01:46:13 -!- puzzlet has joined.
01:46:28 <pikhq> ehird: I usually have two X windows open at a time.
01:46:37 <pikhq> For this purpose, Ratpoison is almost *overkill*. :P
01:46:49 <ehird> ratpoison makes it a real bitch to look at more than one thing at once
01:46:59 <ehird> like, even to have more than one terminal is a pain, imo
01:47:09 <ehird> dwm things are nice because open windows automatically fit and you can hide them if you want
01:48:01 <ehird> my current plan is to build myself a little linux around a tiny, moduleless kernel + static binaries only and dwm...
01:48:14 <ehird> funny how only that and OS X seem to be usable; two extremes
01:48:28 <pikhq> Why the annoyance at dynamic linking?
01:49:05 <pikhq> (the idea, not the implementation; UNIXy things make it hard to actually *build* dynamic libraries)
01:49:08 <ehird> http://blog.garbe.us/2008/02/08/01_Static_linking/ is the most succinct explanation
01:49:23 <ehird> and, to add on to that post: They're just simpler.
01:49:27 <ais523> DJGPP deliberately doesn't have shared objects
01:49:33 <pikhq> Well, they certainly are simpler.
01:49:45 <ehird> http://blog.garbe.us/2008/02/08/01_Static_linking/ gives the non-simplicity nazi reasons :P
01:50:45 <pikhq> It'd be kinda nice if Linux distros weren't so crazy about breaking ABIs.
01:50:54 <ehird> oh, and a.out is simpler than ELF
01:51:05 <ehird> and if you're not using dynamic linking, ELF isn't attractive
01:51:18 <pikhq> Or at least were smarter *about* the treatment of breaking ABIs.
01:51:28 <ehird> simplicity nazi reasons are justified, to be honest; with no kernel modules, static binaries and a.out, I can maintain a distro without pain
01:51:45 <ehird> be nice if someone else did it, of course, but http://stali.suckless.org/ misses the mark ever so slightly :(
01:51:54 <ehird> although they put the kernel in /bin/kernel, which is totally my idea, dammit
01:52:05 <pikhq> ehird: Heck, with no kernel modules, static binaries, and a.out, distro maintainance is little more than compiling things when you upgrade.
01:52:21 <ehird> pikhq: add one to that list — a really simple init system
01:52:31 <ehird> no sysv init, not even bsd init (too opaque file), just two init scripts that call others
01:52:41 <ehird> to activate one, make a binary and add it to the main init file
01:52:44 <ehird> to deactivate, comment out the line
01:52:49 <pikhq> A really simple init system is... Not someething I blame you for.
01:52:53 <ehird> main init files = shell scripts
01:52:58 <ehird> pikhq: tell me about it
01:53:05 <ehird> even adding a simple thing to debian's init.d makes me cry
01:53:35 <pikhq> Make it a couple scripts that run things in /etc/init.d and /etc/halt.d or some such.
01:53:39 -!- ais523 has quit (Remote closed the connection).
01:53:53 <ehird> pikhq: just /etc/init.start and /etc/init.stop, I think
01:53:56 <ehird> no need for tons of runlevels
01:54:02 <ehird> /etc/init.start would look something like:
01:54:16 <ehird> /etc/init/networking/start
01:54:23 <ehird> /etc/init/blahcrapservice/start
01:54:34 <ehird> in fact, forget the directories; /etc/init/networking.start
01:54:38 <pikhq> If you really, truly want it to listen to init signals, make an absurdly simple inittab.
01:54:59 <ehird> eh, does it really matter?
01:55:09 <pikhq> ehird: I was more thinking: for i in /etc/init.d/*; $i ; done
01:55:18 <ehird> pikhq: ah, but what about disabling things temporarily?
01:55:20 <ehird> all the other systems let you do that
01:55:34 <ehird> also, passing arguments
01:55:37 <ehird> for instance, network configuration
01:55:49 <ehird> starting two web servers
01:56:16 <pikhq> for i in /etc/init.d/*; if [ -x $i];then $i;fi ; done
01:56:21 <ehird> /etc/init/httpd.start -c /etc/httpd/site1.conf
01:56:21 <ehird> /etc/init/httpd.start -c /etc/httpd/site2.conf
01:56:44 <ehird> anyway, easy to switch between either.
01:57:00 <ehird> one idea for a package management system:
01:57:26 <ehird> /pkg/kernel/2.6.31.5 is a directory full of directories and symlinks
01:57:40 <ehird> inside, a symlink "kernel" to /bin/kernel
01:57:48 <ehird> and, just making something up
01:57:57 <ehird> a dir usr/, a dir share/, a symlink kernel to directory /usr/share/kerneel
01:58:07 <ehird> so basically, removing a package = rm -r it while following symlinks
01:58:24 <ehird> and it means you can easily view the files in a package, etc
01:59:03 <ehird> that's just an idea, though
01:59:15 <ehird> equivalently is a list of files
01:59:38 <ehird> so i guess that idea is silly, really
02:00:04 <ehird> hey, that's cool; with a list of files, updating a package is just rsyncing the relevant files
02:00:52 <ehird> that + install/uninstall scripts for things like init system entries = package management system
02:01:42 <coppro> what's ehird trying to do? this conversation is tl
02:01:44 <ehird> kinda upsetting when you come up with such a simple solution, knowing the complexity of all the others
02:01:58 <ehird> coppro: the most recentt thing = package management system, aggressively simplified
02:02:51 <ehird> the supertopic = making an incredibly minimalist linux distribution with tiny kernel without modules + static binaries only (and a.out) + init system solely consists of two shell scripts that call other shell scripts and binaries + package management system is so light as to barely exist
02:04:17 <ehird> oh, i just realised an issue
02:04:30 <ehird> if i need to, for whatever reason, use emacs, even for a millisecond... i'd have to create a package for it
02:04:33 <ehird> i'd be a maintainer of emacs!
02:05:33 <coppro> ok, I get you so far - how do you plan to deal with configuration of packages and/or, more importantly, changed configurations?
02:06:14 <ehird> configuration of packages is handled by /etc files. maybe if there is a justification, i will write scripts to handle it, but it'd basically be the same as e.g. gnome's configurators; not needed in the package management system
02:06:41 <ehird> i don't see any issue with this; do you? and, in that context, what would need to be done about changed configurations? I'm not sure what you mean by that term
02:08:01 <ehird> coppro: additionally, i don't see you dealing with configuration of your mom.
02:08:47 <ehird> WELL THAT PUT YOUR RESPONSES OUT OF BUSINESS. so to speak.
02:09:00 <coppro> ehird: well, look at what debian does - it installs a default config in /etc - but some packages offer customization options when installed. This normally isn't particularly necessary, but some packages need configuration options to be useful
02:09:12 <coppro> ehird: and if the user changes the configuration, you need a merge system for upgrades
02:10:32 <ehird> (a) sure, just handle that in whatever global configuration system you desire, be it vi /etc/blah or gnome-whatever — just either have sane defaults if possible (or omit options entirely if it won't work without; just have a commented out template) and, if you must, alert the user in the install script
02:10:32 <ehird> (b) good point, but is this really a big deal? you could copy the default config somewhere, but generally the user's current config will work
02:10:39 <ehird> (and notify the user of its location)
02:10:58 <ehird> if there are backwards incompatible changes, yeah, i guess you should say "HEY USER FIX THIS", but I don't think the existing merge systems can handle syntax changes or whatever
02:13:23 <ehird> coppro: i just don't see merge systems as handling a problem that happens often, and when they do handle it it's either trivial or dissatisfactory
02:13:28 <ehird> i may be totally wrong
02:13:54 <coppro> ehird: debian does a prompt like a version control system in case of a merge conflict
02:14:07 <ehird> Does that address anything I say? If so, what part?
02:14:35 <coppro> <ehird> and when they do handle it it's either trivial or dissatisfactory
02:14:47 <ehird> That was a summary of:
02:14:54 <ehird> [01:10] ehird: (a) sure, just handle that in whatever global configuration system you desire, be it vi /etc/blah or gnome-whatever — just either have sane defaults if possible (or omit options entirely if it won't work without; just have a commented out template) and, if you must, alert the user in the install script
02:14:54 <ehird> [01:10] ehird: (b) good point, but is this really a big deal? you could copy the default config somewhere, but generally the user's current config will work
02:14:54 <ehird> [01:10] ehird: (and notify the user of its location)
02:14:54 <ehird> [01:10] ehird: if there are backwards incompatible changes, yeah, i guess you should say "HEY USER FIX THIS", but I don't think the existing merge systems can handle syntax changes or whatever
02:15:13 <ehird> the last line is the dissatisfactory part, (b) is the trivial part, and (b) is also the rarely-happens part
02:16:45 <ehird> but, really, i'd love for you to prove me wrong; then i can stop being pissed off that everyone else is overcomplicating things so much
02:20:34 <ehird> coppro: you aren't proving me wrong :(
02:21:05 <ehird> THAT'S WHAT THEY A;; SAY
02:26:40 <ehird> should be simple, easy, fun, usable. ha ha ha as if
02:38:16 <pikhq> Gentoo has a script for merging any changes in config files.
02:38:32 <pikhq> If you don't execute the script, it leaves your config files the hell alone.
02:38:34 <ehird> any? lemme guess, a generic merge tool
02:38:43 <ehird> how often is script execution done?
02:38:49 <ehird> i mean, on average
02:39:13 <ehird> I get the feeling that generally, merging is unneeded, and when it's needed it's either really easy to do or a complete renovation, which can't be automated without a lot of pain
02:39:14 <pikhq> I do it every time I upgrade, but that's just me being rather careful & paranoid. How often do most do that?
02:39:20 <pikhq> Uh... Very, very rarely.
02:39:56 <ehird> rightyhothen, won't be needing none of that
02:40:10 <pikhq> Generally when the build hands out a message saying "The format of the config file changed in X manner. Go over your old configuration and make it work."
02:40:27 <ehird> hmm, it's annoying that I can only really optimise for i686
02:40:38 <ehird> because everything above that in the kernel is cpu-specific, right?
02:40:44 <ehird> i.e., not just a "anything not ancient, goddamn"
02:44:59 <ehird> btw: wired ~1.5 Mb/s internet with ~85 ms latency kicks the shit out of 3G.
02:45:28 <ehird> especially since it isn't 15 £ / 1 fucking GiB
02:45:50 <ehird> (currency is too an SI unit)
02:46:26 <pikhq> BTW, if you're going to be doing the static Linux thing, ehird -- glibc sucks ass for static linking.
02:46:37 <pikhq> (namely, it's incapable of doing it properly)
02:46:37 <ehird> you think i'd use glibc?
02:46:46 <ehird> MAYBE I'll use eglibc for programs that absolutely demand it
02:47:03 <pikhq> Eglibc at least fixes the whole "can't do static linking" bit.
02:47:06 <ehird> but glibc? no chance in hell
02:47:20 <ehird> heck, I'm hoping I can just use gcc for the kernel
02:47:30 <ehird> and compile the rest with clang
02:47:34 <ehird> (wonder if clang has baggage if you do static linking)
02:47:48 <ehird> bsmntbombdood: 24" isn't large.
02:47:53 <ehird> lol out of context etc
02:48:01 <pikhq> Well, I hope you don't have a burning itch for something written in C++.
02:48:13 <pikhq> (fortunately, not much that I don't think you'd be using)
02:48:16 <ehird> pikhq: clang does "semi-okay" c++ i believe
02:48:29 <ehird> bsmntbombdood: That's the opposite of what she said!
02:48:36 <pikhq> ehird: Yeah, but I seem to recall that it breaks at least some things.
02:48:41 <ehird> webkit is pretty important.
02:48:47 <pikhq> On the other hand, clang has built Webkit and KDE, so...
02:49:02 <pikhq> Guess those are the two *largest* things anyone will be building that use C++.
02:49:05 <ehird> apparently clang can't compile itself
02:49:26 <ehird> pikhq: latest gcc doesn't do a.out, yeah?
02:49:38 <pikhq> They very recently stopped supporting it, yeah.
02:49:48 <ehird> wonderful, frozen in time
02:49:54 <pikhq> And I'm not sure what LLVM can do.
02:49:57 <ehird> oh well, I'm sure the kernel will compile with it for years
02:50:11 <pikhq> You can still build the kernel with GCC 2, I think.
02:50:25 <pikhq> I know they still *support* GCC 3...
02:50:57 <ehird> pikhq: I was thinking "cool, I can use google's gold linker for speed" but then I realised (a) gold is ELF-only (b) static linking takes like a millisecond
02:51:16 <ehird> llvm uses the platform's linker, I think
02:51:20 <ehird> so a.out should be no trouble
02:51:38 <pikhq> The gold linker supports plugins.
02:51:45 <pikhq> Such as using LLVM for link-time optimization.
02:52:07 <ehird> pikhq: one issue is coreutils type things
02:52:16 <pikhq> So, just add -emit-llvm to your CFLAGS for everything, and you get a lot of link-time optimization.
02:52:21 <ehird> not gonna use gnu coreutils, don't think the bsd ones will work out of the box, and busybox is way too minimal
02:52:26 <ehird> everything else is probably niche and unmaintaine
02:52:39 <pikhq> The NetBSD ones almost certainly will work out of the box.
02:52:39 <coppro> ehird: I believe clang can do the kernel too; not sure
02:52:47 <ehird> i think porting a bsd's and going from there will be the best idea
02:52:55 <ehird> (the BSD coreutils are good, but not perfect)
02:52:59 <ehird> pikhq: interesting
02:53:13 <ehird> coppro: I find that *highly* unlikely; it can compile FreeBSD, though
02:53:23 <ehird> and even that isn't as stable as gcc
02:53:27 <ehird> and this is due to work, iirc
02:53:32 <coppro> ehird: clang's C support is very good
02:53:38 <ehird> I think the clang guys are trying to make the kernel work with it
02:53:39 <pikhq> Linux is very hard to compile; they are very, very GCC-specific.
02:53:42 <ehird> Summary: [META] Compiling the Linux kernel with clang
02:53:59 <ehird> linux is like 40% fucked up gcc bullshit
02:54:10 <ehird> http://llvm.org/bugs/show_bug.cgi?id=4068
02:54:14 <ehird> Created an attachment (id=3486) [details]
02:54:14 <ehird> patch to Linux kernel to build with clang
02:54:18 <ehird> doesn't boot yet though
02:54:21 <pikhq> Clang supports commonly used GCC stuff.
02:54:23 <ehird> and that was only last month
02:54:25 <ehird> and requires patching
02:54:29 <pikhq> Linux's GCC stuff is unique to GCC.
02:54:45 <ehird> pikhq: on the other hand,
02:54:46 <ehird> http://llvm.org/bugs/attachment.cgi?id=2897
02:54:50 <ehird> http://llvm.org/bugs/attachment.cgi?id=3486
02:54:52 <coppro> from the looks of that, they're close
02:54:56 <ehird> is apparently all you need for it to build with clang
02:54:58 <ehird> but booting isn't happening yet
02:55:05 <pikhq> I can't imagine they'd be very far by now.
02:55:06 <ehird> coppro: anyway, it won't be stable for a while, I'm almost sure
02:55:15 <pikhq> Clang has shaped up to be a very good compiler.
02:55:19 <coppro> ehird: they just issued a production-ready release
02:55:21 <ehird> very far from the end, ic
02:55:38 <ehird> so what, it still can't compile a booting kernel
02:55:39 <coppro> not for C++, obvoiusly, but for C and Obj-C
02:55:43 <ehird> and production ready for X != production ready for kernel
02:55:50 <ehird> also, webkit and kde can build with it, apparently
02:55:53 * pikhq wonders if llvm-gcc can build a functioning kernel
02:55:56 <coppro> you just said it wouldn't be stable, and I wanted to disagree!
02:55:57 <ehird> good enough for me as far as C++ goes
02:56:00 <ehird> pikhq: click the damn bugzilla link
02:56:04 <ehird> it needs two more patches than clang
02:56:08 <ehird> coppro: not stable FOR THE KERNEL
02:56:16 <ehird> or do you think as soon as it boots it'll be a stable kernel
02:56:17 <pikhq> ehird: I just started.
02:56:26 <ehird> if so, um, have you ever read the kernel source
02:56:47 <ehird> I'm just worried that statically linked llvm binaries will be huge
02:57:01 <pikhq> Why would they be?
02:57:16 <ehird> support functions and stuff; I just doubt it's really been tested
02:57:35 <ehird> I wonder how big a total llvm+clang toolchain is vs binutils/gcc
02:58:03 * coppro could try to compare, but he's lazy
02:58:04 <pikhq> At least judging from my experience building llvm+clang and binutils/gcc, llvm+clang is at least signficantly faster to *compile*.
02:58:21 <coppro> yes, llvm+clang is far faster to build and to execute
02:58:42 <coppro> and doesn't have a hobbled-together artificial type system
02:58:48 <ehird> which is a relief; I don't expect to provide a kde package, but webkit releases are, I think, quite often
02:59:24 <coppro> you could use EDG to compile the kernel :P
02:59:26 <ehird> i'll probably maintain the distro by having a huge tree with every package's source, install/uninstall files and any additioanl files to install, rsync, and mk
02:59:35 <coppro> just to say you don't use GCC
02:59:44 <ehird> just have a huge mkfile that includes all the others (which I'll craft myself)
02:59:49 <pikhq> Hmm. So, llvm-gcc doesn't build Linux because Linux relies on .code16gcc.
02:59:53 <pikhq> Holy crap. That's awful.
03:00:02 <ehird> of course, the package mkfiles will do "make clean"
03:00:04 <ehird> for the actual packages
03:00:12 <ehird> so each package upgrade is a clean build
03:00:19 <pikhq> .code16gcc is the most crazy hack in GCC.
03:00:23 <ehird> but it'll handle only building packages i've updated the sources or additional files
03:00:29 <pikhq> And GCC supports a *lot* of crazy hacks.
03:00:50 <ehird> pikhq: i wish the bsds weren't so kernel-userspace wed, or I'd be using one :P
03:01:27 <ehird> I wonder how quickly WebKit compiles with clang (excluding link time, since, you know, static binaries/libraries)
03:01:35 <ehird> hooray for .a (does anyone even remember using .as recently?)
03:02:31 <pikhq> ehird: Well, you could check to see what Debian is doing to have a GNUy userspace on BSD. I can't imagine it being *that* crazy.
03:02:37 <pikhq> On the other hand, it is Debian...
03:02:46 <ehird> well, sure, you can compile just the kernel
03:02:53 <ehird> the point is that the bsd teams maintain the userspace AND the kernel
03:03:02 <coppro> ehird: Roughly 50% of the time on average I think
03:03:05 <ehird> so the kernel will be developed according to the userspace
03:03:19 <ehird> coppro: what's that in response to?
03:03:28 <coppro> ehird: you wondering how long WebKit takes to compile
03:03:41 <coppro> though GCC can be massively sped up with precompiled headers
03:03:42 <ehird> 50% of time as with gcc, that is?
03:03:47 <ehird> i don't know how long it takes with gcc, really :)
03:03:57 <ehird> 30 minutes? two hours?
03:04:01 <coppro> Clang currently doesn't support precompiled headers for C++ and it's not on the todo list
03:04:18 <coppro> ehird: I'd guess ~35, but that's pretty random and arbitrary
03:04:40 <ehird> so ~17.5 minutes with llvm/clang
03:04:43 <ehird> does that include link time? I imagine so
03:04:53 <ehird> linking large shared libraries takes like 5 years
03:05:20 <ehird> hmm, that's a good point; anyone know any a.out linkers that can be compiled without dynamic linking support and aren't made by gnu?
03:05:45 <ehird> hmm, llvm-ld can link?
03:05:50 <ehird> only llvm things, obviously
03:06:01 <ehird> The llvm-ld program has limited support for native code generation, when using the -native or -native-cbe options. Native code generation is performed by converting the linked bitcode into native assembly (.s) or C code and running the system compiler (typically gcc) on the result.
03:06:42 <pikhq> gold supports LLVM and native linking...
03:06:51 <ehird> gold also only supports ELF.
03:06:58 <pikhq> That's kinda lame.
03:07:05 <ehird> Well, that's why it's fast.
03:07:16 <ehird> And speed optimisations don't really matter when everything is static...
03:07:22 <ehird> Since a static linker is, you know, really trivial.
03:07:36 <ehird> Like three steps removed from cat.
03:07:52 <pikhq> Gold supports doing optimisation at link time *instead* of compile-time, FWIW...
03:08:08 <ehird> Yes; seems like a lot of complexity for little gain.
03:08:12 <ehird> I wonder how much stuff assumes /usr exists.
03:08:21 <pikhq> Though, so does llvm-ld.
03:08:28 <pikhq> Just a bit more pain to get working.
03:08:37 <ehird> llvm-ld doesn't actually make a native code executable, though
03:08:40 <ehird> at least, from what i gather
03:08:43 <ehird> it makes an llvm bitcode file
03:09:07 <pikhq> And you can then compile that. Yeah. That's what makes it a pain.
03:09:23 <ehird> Yes. Beats the pain of using gcc, though.
03:09:36 <ehird> In fact, a regular user development system won't include gcc, hopefully.
03:09:36 <coppro> and makes compiling for different platforms easy
03:09:45 <bsmntbombdood> xscreensaver-command: no screensaver is running on display :0.0
03:09:46 <ehird> Of course, you can install it if you want to compile an odd program that requires it.
03:10:13 <ehird> In fact, I wonder if there's even any GNU software a typical system willl run...
03:10:34 <ehird> bsmntbombdood: so run two
03:10:50 <coppro> ehird: you could distribute packages in llvm bitcode and have the package system build them as they arrive - it's a short step and would make things 1000000000x easier
03:10:56 <bsmntbombdood> xscreensaver: 20:10:50: already running on display :0.0 (window 0x1c00002)
03:11:30 <ehird> coppro: But it shouldn't matter what's LLVM and what's not. Currently, my package manager's updating/installing process would consist of rsync and running a shell script.
03:11:36 <ehird> Why would that help, iincidentally?
03:11:44 <ehird> bsmntbombdood: ah, bad software
03:11:49 <ehird> lemme look at xscreensaver's site
03:11:51 <coppro> ehird: then you don't need to deal with ~8 platforms to build on
03:11:55 -!- BeholdMyGlory has quit (Remote closed the connection).
03:12:00 <ehird> coppro: ah, uh, i'm meant to support non-x86?
03:12:04 <ehird> nobody informed me of this :)
03:12:19 <ehird> coppro: well, I'm only targeting desktop machines and maybe servers
03:12:20 <ehird> besides, I'd still need the infrastructure for binaries
03:12:31 <ehird> so it basically comes down to disk space; the bitcode format adds more complexitty
03:12:36 <Sgeo> Is http://www.geocities.com/Athens/Oracle/9941/index.html still up for anyone but me?
03:12:39 <ehird> thankfully, migrating to that should be easy
03:12:53 <coppro> you could also make the server do the bitcode compilation ondemand - even more complex though
03:13:31 <pikhq> coppro: That's a stupid idea, anyways.
03:13:32 <ehird> ugh, the xscreensaver faq says nothing
03:13:43 <pikhq> The bitcode generated by C compilers is target-specific.
03:14:06 <ehird> then yes, that is a stupid idae
03:14:12 <coppro> fine then, store output in llvm assembly code
03:14:14 <pikhq> Because sizeof( ) is done by the C compiler.
03:14:31 <pikhq> Oh, and *everything* that implies...
03:14:48 <ehird> store the output in llvm assembly? that still has sizeof() issues
03:14:52 <ehird> and is leaning further to a source distro!
03:14:59 <pikhq> That would at least work if you were dealing with a language that's not C-esque.
03:15:01 <coppro> ehird: yeah, I said that before he said his objection
03:15:22 <coppro> now, it's worth pointing out if the entire system were built that same way, it would work
03:15:27 <ehird> heh, amusingly, turning it into a source distro would be easy enough
03:15:38 <ehird> point the package manager to a local source instead of an rsync://
03:15:47 <ehird> and just set up the package build environment
03:15:50 <ehird> + download aaaaaaaaaaaall of it
03:15:57 <ehird> of course, keeping the source in sync is then your responsibility
03:16:02 <ehird> and like hell will you get any support...
03:16:09 <bsmntbombdood> upgraded xscreensaver and now it works properly :P
03:16:13 <coppro> ehird: you wouldn't give support in any place
03:16:18 <ehird> *bsmntbombdood: like magic!
03:16:34 <coppro> ehird: your history of being a jerk
03:16:34 <ehird> i'm perfectly happy to help people who don't do stupid stuff to their system
03:16:49 <ehird> plenty of other people are jerks and still offer support
03:17:30 <ehird> bsmntbombdood: Xsession or something
03:17:42 <coppro> your wm may also have an autostart system
03:17:59 <ehird> coppro: anyway, as long as you don't pretend you know more about the problem than you do and didn't do something really stupid, I'm happy to offer civil help... I'm not always a jerk, you know :P
03:18:05 <ehird> also, bsmntbombdood uses ion
03:18:09 <ehird> i somehow doubt it has such a thing
03:18:52 <ehird> hmm, forgot to include "install dependency packages" in my mental model of a package install
03:19:20 <pikhq> Eh, not necessarily needed.
03:19:28 <pikhq> Slackware seems to do just fine without it.
03:19:31 <ehird> should probably dump a really trivial dependency map elsewhere, too, so that i can remove unused packages
03:19:40 <ehird> pikhq: yeah, but installing packages in slackware is a bitch.
03:20:05 <coppro> I've got a new bad idea
03:20:25 <coppro> how about a distro where there's no packages and every program brings its dependencies along with it?
03:20:27 <coppro> wait... that's Windows
03:20:32 <ehird> simple way to solve the "the autoremover thing wants to remove packages I explicitly installed!" thing: a fake package, say _explicit or something, depends on every package you explicitly install
03:20:40 <ehird> (and is a null package otherwise; not on disk, just in the dependency map)
03:20:43 <pikhq> coppro: Been done.
03:20:48 <ehird> bonus: lets you ask it what packages you explicitly installed
03:20:53 <ehird> coppro: that's not windows, that's OS X
03:20:58 <ehird> and it works quite well, fwiw
03:21:10 <ehird> like, it's the same thing as static binaries, really
03:21:12 <coppro> how is that not windows?
03:21:22 <ehird> windows shit often puts stuff in shared directories and the like
03:21:28 <ehird> whereas OS X things almost universally have it all inside the .app bundle
03:21:35 <pikhq> Windows puts stuff in shared directories haphazardly.
03:21:41 <pikhq> Without regard for breaking anything.
03:21:43 <coppro> I didn't say it necessarily kept those dependencies to itself
03:21:50 <pikhq> OS X has almost everything in the .app bundle.
03:21:57 <ehird> coppro: well, then that is of course a horrid idae
03:22:08 <pikhq> And Linux assumes that what it needs has been installed already.
03:22:17 <pikhq> (well, or makes sure it gets installed)
03:22:39 <ehird> proprietary software on linux is a bitch
03:22:40 -!- linf has joined.
03:22:49 <ehird> thankfully, that's pretty much... mathematica and games
03:22:52 <coppro> the independent dependency model is fine as long as disk space is not a concern
03:22:59 <ehird> linf: ha, we caught you
03:23:09 <ehird> linf: Mr Russian music... is in #tokipona, and is n=nikita
03:23:10 <coppro> ehird: and mathematica isn't needed any more :P
03:23:11 <pikhq> coppro: It at least works, yes.
03:23:16 <ehird> linf: IS THAT YOU, LAMENT
03:23:22 <ehird> you trolling scoundrel, you
03:23:26 <ehird> coppro: replaced by what
03:23:48 <ehird> lament: he was in here ages ago, being oblivious to the purpose of this channel and beingn incomprehensible.
03:23:59 <ehird> I guess the .ru should have tipped off that it isn't you, though
03:24:06 <ehird> but n=nikita + #esoteric + #tokipona seemed a fair beet
03:24:14 <lament> wow, that's pretty cool
03:24:48 <ehird> lament: he doesn't actually know what this channel is for, I think he tried talking about russian folk music or something before, but i couldn't tell. admittedly it might be another linf
03:24:51 <ehird> but i doubt _that_
03:25:01 <linf> lament: hahahahaha
03:25:03 <ehird> coppro: sure, for the pure mathematics stuff
03:25:08 <ehird> mathematica is still useful for some things though
03:25:13 <ehird> plus, more importantly, it's fun
03:25:36 * coppro has not really used mathematica
03:26:00 <ehird> e.g. opening up images, applying various transformations, hooking it up into stuff from the web, then doing statistics on them and the like
03:26:10 <ehird> just a sort of "unified analysis and munging environment"
03:26:17 <coppro> mathematica does that? O_o
03:26:18 <ehird> quite entertaining with the right data set
03:26:29 <ehird> coppro: yes, and it also connects to wolfram's servers to download a ton of data really fast
03:26:34 <ehird> e.g. historical gdps, and the like
03:26:38 <ehird> in a tiny function call
03:26:52 <ehird> i dislike wolfram, and mathematica is really, really slow; but it's an awful lot of fun
03:27:23 <ehird> lament: btw, did linf just find your name hilarious?
03:28:56 <ehird> hmm... proprietary graphics drivers are usually modules, right?
03:29:04 <ehird> but you can link module binaries into the kernel, if i'm not mistaken
03:29:21 <coppro> you can compile modules in or load them at runtime
03:29:31 <ehird> I explicitly want to leave out the module support
03:29:45 <ehird> the runtime loading, that is
03:29:55 <ehird> since they're basically dynamic libraries, except in the kernel, which is an even less fun prospect than in userspace.
03:30:10 <coppro> oh right, you're mr. static
03:30:23 <ehird> for good reason! i'm welcome to hear arguments for kernel modules.
03:30:26 <coppro> fine, as long as you include fuse
03:30:51 <ehird> fuse is a bit slow
03:31:12 <ehird> by a bit i mean really
03:31:15 <coppro> you don't use fuse for normal filesystems
03:31:21 <coppro> you use fuse for stuff like mounting an ftp server
03:31:53 <ehird> (which pisses me off; there's no good way to have a proper ntfs driver in the kernel nowadays)
03:32:41 <ehird> coppro: anyway, if fuse doesn't increase the kernel size too much, I'll stronly consider it
03:32:43 -!- calamari_ has joined.
03:32:45 -!- FireFly has quit ("Later").
03:32:46 -!- linf has left (?).
03:33:06 <ehird> (although I do wonder if there's a 9P mounter thingy for the kernel)
03:33:09 -!- FireyFly has quit (Read error: 54 (Connection reset by peer)).
03:33:12 <ehird> that'd certainly be simpler
03:33:15 <coppro> hmm... you know what would be cool
03:33:17 <pikhq> The kernel has 9P support.
03:33:25 <pikhq> I don't know how good it is, but it does have 9P support.
03:33:31 <ehird> coppro: a pony? totally
03:33:35 <ehird> pikhq: huh, really?
03:34:05 <ehird> i wonder if anyone's made an X11-compatible thing that's better than X yet. ha ha, only joking.
03:34:07 <coppro> /usr/bin is a fuse filesystem where attempting to run an executable that isn't installed automatically prompts to install, installs if requested, and then runs
03:34:17 <coppro> ehird: isn't that by definition impossible?
03:34:38 <ehird> coppro: you'd think so, but the current Xorg is really terrible
03:34:44 <ehird> you could, at least, majorly clean it up and have minor functionality imprrovements
03:34:53 <ehird> should be possible to debloat Xorg, though. with enough effort. sigh
03:34:57 <pikhq> The amazing thing about Xorg is that it's an improvement over the past.
03:35:01 <coppro> ehird: then you should have said Xorg, not X
03:35:09 <ehird> yeah yeah whatever
03:35:25 <ehird> maybe I'll just only compile in the new-fangled thingy that only the unstable intel drivers support
03:35:32 <pikhq> (amazing and very, very scary)
03:35:37 <ehird> after all, the proprietary ATI driver doesn't support old cards at all nowadays
03:35:43 <ehird> that is — cards that were commonly used IN 2006
03:36:01 <ehird> coppro: You stab fglrx, but have you used the open source drivers?
03:36:07 <ehird> So. Amazingly. Slow.
03:36:14 <coppro> ehird: yeah, but they were stable
03:36:29 <ehird> Yeah, stable, but the jagged artifacts when moving a window slowly...
03:36:33 <ehird> Horrible. Just horrible.
03:36:40 <coppro> you can always disable compositing
03:36:49 <ehird> It was the same regardless.
03:36:54 <ehird> That's how slow they are.
03:36:55 <coppro> nothing that bad when I used it
03:37:12 <ehird> coppro: also, no way in hell will i use xfree86
03:37:21 <coppro> fglrx would hang if I attempted to run anything as difficult as glxgears
03:37:25 <ehird> the license change shows them to be complete buffoons. plus it's totally dead and there aren't any drivers for it etc.
03:37:30 <coppro> granted, half the time the open-source ones would simply give up at glxgears
03:37:46 <coppro> but at least they didn't randomly cause me to have to reboot
03:37:56 <ehird> coppro: fglrx DROPPED SUPPORT for my card from 2006. and I couldn't use an old version. Why? Because the only version supporting the new X11 version doesn't support my card.
03:38:09 <ehird> so, ATI to users of my card: Buy a new card or fuck off.
03:38:31 <ehird> At least they released their hardware specs.
03:38:59 <ehird> nvidia have better proprietary drivers, but ATI support the OSS drivers more
03:39:27 <Ilari> Jagged artifacts? Bottom and top of window not moving in sync?
03:39:38 <ehird> Ilari: basically the whole window out of sync, yeah
03:39:53 <ehird> for windows of any size, even dialogs
03:40:40 <ehird> incidentally, anyone have any strong opinions on the default (and "official") FS to use?
03:40:58 <coppro> I'm a fan of ext though
03:40:59 <Ilari> The bottom and top not being in sync is caused by lack of syncing the updates to Vretrace.
03:41:28 <ehird> I'm leaning towards JFS; it doesn't seem to be hugely actively developed, but it's not abandoned, and it's like XFS, except it doesn't like to lose data on crashes, and is way faster at metadata (including creating files, etc) operations, making it one of the fastest "real" linux filesystems
03:41:46 <ehird> and the fsck takes *3 seconds* — and that's when it has to recover some data
03:41:50 <ehird> (and it recovers very well)
03:41:57 <ehird> it only takes 2 seconds for less data to recover
03:42:04 <ehird> and presumably almost nil time if there's nothing to do
03:42:57 <ehird> I'm pretty much not considering ext because I hate the multiple-minute fsck times more then the plague
03:43:23 <Ilari> And apparently JFS supports SELinux...
03:43:37 <ehird> not really planning to use selinux, though
03:43:50 <ehird> way too much fuss, too little gain; save the capability-based security for ehirdOS
03:43:52 <coppro> ext fsck isn't usually very long if there's no corruption
03:44:00 <ehird> coppro: if you have a tiny disk
03:44:01 <coppro> if there's problems, though...
03:44:06 <ehird> it takes about 3 minutes, all the time, for me
03:44:16 <ehird> and longer if it has to fix stuff
03:44:28 <ehird> compared to that, 3 seconds when the system hard reset while doing stuff is godly
03:44:56 <coppro> how big is your disk?!
03:45:16 <ehird> um, this one? 250 GB
03:45:40 <ehird> though the 3 minute figure is from the old PC, which has a 500 GB disk (and partition)
03:45:48 <ehird> since I don't use linux on this, at least not for extended periods of time
03:46:08 <ehird> it had a consumer-level Athlon when doing that, though, and 2 GiB of RAM
03:46:15 <ehird> so it's definitely the fsck being the bottleneck
03:46:19 <ehird> (and the 500 GB disk is 7200 rpm)
03:47:59 <Ilari> If windows get severe jagged atefacts when moving, the video drivers must REALLY suck. I can only get some artefacts on this computer, video driver is pretty much the suckiest (fb) and hardware sucks hard.
03:48:18 <ehird> yeah; it's the non-radeonhd (radeonhd doesn't work) open source one
03:48:21 <ehird> called ati or radeon
03:48:31 <coppro> ehird: this is 250GB, but it is partitioned
03:48:40 <ehird> hardware is a radeon x1600, like a mid-range notebook card from 2006
03:48:50 <ehird> releassed in 2005 i think
03:49:07 <ehird> Ilari: i mean, it's not hugely severe, but if you look at the edges you definitely notice it every time
03:49:17 <Ilari> Heh... Moving window slowly uses something like 50% CPU.
03:50:29 <ehird> Tasty like faeces!
03:50:49 <coppro> Homer's second-best line!
03:51:18 <Ilari> One useful use of framebuffer. When X craps out (like if its keyboard driver(!) craps out), ALT-SYSRQ-R, ALT-F1.
03:51:29 <ehird> eh, I was going to quote The Odyssey
03:53:26 <Ilari> Yes, had keyboard driver of Xorg crap out... Without it working, CTRL-ALT-Fx doesn't work.
03:54:13 <ehird> hey, has xorg fixed graphics drivers so that if it craps out X doesn't die yet?
03:55:12 <coppro> Ilari: Alt + SysRq + R
03:55:31 <ehird> coppro: that's, like, the best thing about windows vista. not that that sets the bar terribly high
03:56:49 <ehird> there should be a usb peripheral for linux that's a magic wand like the wii remote thing, right, and you can do gestures to do magic sysrqs
03:56:53 -!- Gregor has quit (Read error: 145 (Connection timed out)).
03:56:56 <ehird> hard-rebooting becomes merely a spell.
03:58:39 -!- Gregor has joined.
03:58:59 <ehird> bsmntbombdood: dwm is the best of breed of the xmonad-types
03:59:04 <ehird> well, dwm-types really
03:59:37 <ehird> very simple, no retarded the-config-file-is-the-program-also-it-depends-on-ghc-because-fuck-you-we-like-haskell-config stuff, and yeah.
04:00:59 <ehird> NEW dwm creates a view for each Xinerama screen.
04:01:04 <ehird> so it should work well with multihead setups.
04:09:40 -!- bsmntbombdood has quit (Remote closed the connection).
04:12:18 * pikhq notes that "the-config-file-is-the-program" would work much much better if Haskell had an eval function...
04:12:52 <pikhq> Though the closest you get to eval in Haskell is linking against GHC so you can use GHC's libraries to create a (small) interpreter.
04:13:09 <ehird> xmonad just compiles the config file
04:13:09 -!- Azstal has quit (Read error: 110 (Connection timed out)).
04:13:20 <pikhq> Yes. I'm well aware.
04:13:22 <ehird> you know, I thought I knew C fairly well until I looked at some suckless code
04:13:30 <ehird> i think i'm not dumb enough
04:13:36 <pikhq> Should I read some?
04:13:36 <ehird> like, my mind creates too many abstractions ahead of itself
04:14:05 <ehird> pikhq: Absolutely. http://dl.suckless.org/dwm/dwm-5.7.2.tar.gz is dwm; dwm.c is 2018 lines long in total.
04:14:10 <ehird> and the only other code file is config.h
04:14:36 <pikhq> ehird: That's very impressive considering a) it's C b) it's C that *uses X11*.
04:14:39 <ehird> well, it's config.def.h technically, but you copy it to config.h and modify.
04:14:47 <ehird> pikhq: Absolutely.
04:15:07 <ehird> "To understand everything else, start reading main()." it gives a basic overview of some stuff before that, but I love that concept
04:15:11 <ehird> like, it's uber-procedural code
04:15:16 <ehird> so procedural that you can just follow from main to understand it
04:15:17 <pikhq> Also, this appears to be designed such that a straight Makefile is *actually sane*.
04:15:41 <ehird> they even make their manpage look easy :(
04:16:34 <pikhq> ... Holy crap that's nice C.
04:16:49 <pikhq> They make C look easy.
04:16:54 <pikhq> THEY MAKE C LOOK EASY.
04:17:07 <ehird> that was my reaction too :/
04:19:26 <pikhq> Is this... Is this what UNIX was meant to be?
04:19:54 <bsmntbombdood_> the very top, middle pixels are not visible on the monitor
04:21:14 <ehird> pikhq: the other programs whose code i think is very representative of suckless style are dmenu (menu selection/completion system; designed to be used to e.g. launch programs with dwm; http://dl.suckless.org/tools/dmenu-4.0.tar.gz) and ii (filesystem-based irc client; http://dl.suckless.org/tools/ii-1.4.tar.gz)
04:21:49 * pikhq would like to see suckless create a C compiler.
04:21:55 <ehird> well, sic is too, but it's a terminal-based IRC client in less than 250 lines, so it's not surprising that it's very simple
04:22:03 <ehird> pikhq: I think they'd point you to plan9's 9c :P
04:22:08 <pikhq> I'm imagining something similar to tcc.
04:22:13 <pikhq> Oh, right, plan9's C code.
04:22:18 <pikhq> That, too, is pretty nice.
04:22:55 <pikhq> It helps that they get a library better than libc to work with. ;)
04:23:06 <ehird> dmenu, I think, has their largest seeming-task-complexity : code-simplicity gap
04:23:16 <ehird> i.e., it's the code that seems the most like it should be more complex
04:24:37 <pikhq> "Many (open source) hackers are proud if they achieve large amounts of code, because they believe the more lines of code they've written, the more progress they have made." -- Suckless manifesto.
04:25:05 * pikhq notes that lines of code is more of an inverse metric.
04:25:40 <ehird> another good metric: count of data structures
04:25:43 <ehird> and size of data structures
04:25:56 <ehird> here's an example bot written for ii, incidentally:
04:25:58 <ehird> tail -f \\#<CHANNEL>/out |
04:25:58 <ehird> while read foo ; do
04:25:58 <ehird> name=$(echo $foo | awk '{print $2}' | sed 's,<\\(.*\\)>,\\1,')
04:25:59 <ehird> if awk 'BEGIN{srand(); exit rand()<.1)}' ; then
04:25:59 <ehird> echo "$name: WHAT??" ;
04:26:07 <ehird> substitute <CHANNEL>, obviously
04:26:27 <ehird> (and add #channel/in, too)
04:26:35 <pikhq> Heck, that's one of the things I like about Haskell: your code is succinct and can be very simple.
04:26:55 <ehird> haskell code tends to be complex because the environment around it isn't haskelly
04:27:05 <ehird> for algorithmic work, though, it's very simple
04:27:35 <ehird> oh man, I wonder how big static binaries made with gcc are
04:27:39 <ehird> probably at least 20 MiB
04:27:52 <pikhq> Depends on the libc.
04:28:00 <pikhq> If it's glibc, "very large".
04:28:41 <ehird> that's a point, I should pick a libc
04:28:55 <ehird> bsmntbombdood_: define ths
04:29:06 <bsmntbombdood_> < bsmntbombdood_> the very top, middle pixels are not visible on the monitor
04:29:14 <ehird> define not visible
04:29:17 <ehird> covered by bezel? dead?
04:29:39 <ehird> your monitor manufacturer doesn't care about you
04:29:47 <ehird> so they decided to make it look shinier instead.
04:30:24 <ehird> Which sort is this?
04:30:49 <pikhq> Small program I wrote a few years back when I was just figuring out sorting algorithms in C.
04:30:51 <ehird> If it's GNU, then it gratuitously uses half of the standard library just because it can. Also, glibc is huge.
04:30:59 <ehird> right, static gnu sort would be way bigger
04:31:07 <ehird> I blame glibc, then
04:31:17 <pikhq> Glibc is freaking huge, yes.
04:31:44 <ehird> hmm... looks like my libc options are dietlibc or newlib
04:32:02 <ehird> Well, that's really embedded-targeted, isn't it?
04:32:06 <ehird> newlib is intended for embedded systems, but cygwin uses it as its libc
04:32:17 <ehird> and dietlibc, while mainly used on embedded devices, is optimised for small size
04:32:28 <ehird> (and omits uncommonly used functions)
04:32:53 <pikhq> uclibc's embedded support consists of the ability to disable uncommonly used functions.
04:32:58 <pikhq> And working on uclinux.
04:33:46 * ehird looks up newlib's license
04:33:54 <ehird> newlib is tempting because Cygwin uses it
04:34:05 <ehird> and Cygwin works with, well, not a lot, but that's Cygwin's fault
04:34:45 <ehird> well, it's also GPL, but the lib itself is LGPL, I gather
04:35:10 <ehird> dietlibc is... wait, what? GPL?
04:35:15 <pikhq> Newlib's main tempting property is that it has been shown that many things can *use* it.
04:35:41 <ehird> dietlibc is gpl, and it's intended only for static linking
04:35:45 <ehird> doesn't virility apply?
04:36:18 <ehird> that suxx big butt
04:36:26 <ehird> pikhq: heh, it's the readline argument, isn't it?
04:36:36 <ehird> what if I linked a third party program against it?
04:36:40 <ehird> are they violating the GPL?
04:37:06 <pikhq> It's... Very fucking irritating to resolve.
04:37:28 <ehird> yeah, well, copyright is fucked.
04:37:44 <pikhq> The LGPL's behavior in such cases is at least well-defined.
04:37:53 <ehird> but i don't want to play with fire. i don't believe in intellectual property, but that's for personal stuff
04:38:13 <pikhq> Then I guess dietlibc's out.
04:38:14 <ehird> such a shame, because dietlibc looks rather good
04:38:31 <ehird> oh, and it has a subsystem named libcruft
04:39:06 <ehird> and the code looks nice — I should stop before I get annoyed that it's GPL
04:39:35 <pikhq> uclibc is also apparently usable for 'most things'.
04:39:48 <ehird> Q: GPL sucks! Now I can't compile my BSD programs with the diet libc!
04:39:49 <ehird> A: Wrong. You can compile them, and you can use them. You just can't
04:39:49 <ehird> redistribute the binaries. If you are a distribution vendor and want
04:39:49 <ehird> to use the diet libc to make BSD licensed binaries for the install
04:39:49 <ehird> or rescue floppy which you sell commercially, please talk to me.
04:40:05 <ehird> eh, blow me, i don't give a fuck about your anti-commercial bent, i just want to make a distro
04:40:25 <pikhq> ... Even Stallman doesn't support using the GPL for such a case.
04:40:41 <pikhq> More extreme than Stallman regarding free software = *facepalm*
04:40:48 <ehird> Q: Can I compile or use the diet libc with a compiler that is not gcc?
04:40:48 <ehird> A: Compile: no. Use: yes.
04:40:48 <ehird> ALSO, BLOW ME AGAIN
04:40:52 <ehird> admittedly probably most are like that
04:41:00 <ehird> does it even make sense to compile a libc with clang/llvm?
04:41:20 <pikhq> I don't see why not.
04:41:29 <ehird> llvm does have some overhead, yeah?
04:42:04 <ehird> So, it's uClibc vs newlib.
04:42:14 <pikhq> The only "overhead" is that LLVM "produces very slightly slower" code than GCC.
04:42:15 <ehird> In computing, uClibc is a small C standard library intended for embedded Linux systems. uClibc was created to support uClinux, a version of Linux not requiring a memory management unit and thus suited for microcontrollers (uCs; the "u" is a romanization of μ for "micro").[2]
04:42:24 <ehird> so, clearly uClibc's development focus will be embeddedness
04:42:33 <pikhq> (GCC has some optimisations that LLVM doesn't)
04:42:44 <ehird> LLVM produces better code in some cases, though.
04:42:50 <ehird> Or at least, clang.
04:43:02 <ehird> 5 May 2009, SVN -> GIT
04:43:02 <ehird> We've migrated from SVN to GIT. SVN is frozen read-only before the conversion, so check out the Developing links and such for updated instructions.
04:43:02 <ehird> well, the uclibc guys are modern...
04:43:09 <ehird> newlib is a red hat project with all that entails
04:43:23 <ehird> "uClibc++" hey that's nice, a C++ lib too
04:44:04 -!- Slereah has quit (Read error: 104 (Connection reset by peer)).
04:44:17 -!- Slereah has joined.
04:44:22 <ehird> "The GNU C library is a great piece of software, make no mistake." —uclibc FAQ
04:44:49 <ehird> "So uClibc is smaller then glibc? Doesn't that mean it completely sucks? How could it be smaller and not suck?"
04:44:49 <ehird> i think i'm going to start calling such things "strawman FAQs"
04:45:35 <pikhq> Still, uclibc seems to function well for a system libc.
04:45:48 <ehird> the faq is very embedded-oriented, it seems
04:45:56 <ehird> and it seems that uclibc is gcc-only
04:46:04 <ehird> that is, you have to build uclibc-using programs with gcc.
04:46:31 <pikhq> So's many things that clang builds just fine. ;)
04:46:44 <ehird> true, but the gcc is *patched*
04:46:47 <pikhq> clang supports most of GNU C, you know.
04:46:52 <ehird> you need a patched gcc to build uclibc binariess.
04:46:58 <ehird> even just binaries that use it
04:47:00 <pikhq> Never mind. newlib it is!
04:47:53 <ehird> Plus eglibc for broken programs.
04:48:00 <ehird> (STRFRY() OLOLOLO)
04:50:44 <ehird> this should be easy, then.
04:50:54 <ehird> one thing I'm unsure of how to handle is permutations of kernel configs
04:51:06 <ehird> the main distros usually solve this by a whole shitload of modules, plus a really large base kernel config, but that sucks
04:52:09 <pikhq> Very large kernel or "build one".
04:52:20 <pikhq> (which sucks less, I wonder...)
04:53:15 <ehird> methinks I'll just include what people actually use as far as drivers go, plus perhaps separate kernel packages for graphics drivers, as they're the only big proprietary kind of thing, everything else should work ootb
04:53:29 <ehird> otherwise, build one, it should be quite easy with the package system
04:54:00 <ehird> pikhq: Think I should use TuxOnIce? It's one of the hibernate/restore systems; apparently it can hibernate and restore in just seconds, which sure as hell beats the S2 stuff.
04:54:07 <ehird> Fedora uses it, I believe.
04:54:28 <ehird> Thing is, you have to patch the kernel.
04:55:31 <ehird> http://www.youtube.com/watch?v=k_T4ZcPkqVA it's still quite slow, though...
04:55:51 <ehird> coming from an OS X background where it takes about 3 seconds either way
04:56:02 <ehird> 'cause it uses a hybrid suspend/hibernate
04:58:00 <ehird> wonder how big my kernel will be
04:58:09 <ehird> i bet around 2 MiB
05:00:15 <ehird> okay, so that's most things sorted
05:01:02 <ehird> of course, I need to find a decent linux system from which to develop it...
05:08:17 <ehird> Doo doo doo, ood ood ood
05:08:33 <ehird> Clearly I should use LFS to build it! No. No.
05:10:50 * ehird looks up what mailing list software suckless.org uses
05:11:13 <ehird> mlmmj. "maling list management made joyful". Cute name.
05:11:32 <ehird> It didn't make me rage when I wanted to post to it without receiving replies, so that's good.
05:21:20 <ehird> come to the dwm side! we have cookies. uh, actually, we don't. we have... lots of code? hmm. we don't have much actually
05:21:29 <ehird> i have this stick if you want it.
05:21:38 <ehird> it's a good stick!
05:21:57 <fax> haskell: gives you the willies
05:22:12 -!- bsmntbombdood has joined.
05:22:25 <ehird> nothing wrong with haskell, something wrong with using it as a configuration language thouh
05:31:14 -!- Oranjer has joined.
05:31:38 <Oranjer> a donkey shaved as one shaves a poodle?
05:31:51 <ehird> a donkey that is a poode
05:32:06 <ehird> "skvm is a lightweight volume manager for GNU/Linux. It depends on hal and dbus."
05:32:13 <ehird> why is this on the suckless code server
05:32:15 <ehird> i mean hal and dbus, really?
05:32:28 <ehird> i'm not talking to you
05:33:39 <ehird> fax: yes your momm.
05:34:55 <pikhq> I could've sworn HAL sucked sufficiently that most everyone was trying to replace it...
05:35:31 <ehird> and dbus is pretty heavyweight when you can just use more lightweight IPC
05:35:34 <ehird> by pretty i mean really
05:35:50 <ehird> all it needs now is to require udev too :-P
05:36:40 <pikhq> Dbus is meant work sanely with GNOME and KDE... So, yeah, it's obviously going to be pretty heavyweight.
05:39:20 <ehird> maybe i need to write an email client that sucks less, since all of them have such retarded approaches to threading
05:39:44 <pikhq> I could've sworn that threading in general had retarded approaches. ;)
05:39:54 <ehird> not that kind of threading. :P
05:40:17 <Oranjer> well, ehird, what's your approach to threading?
05:40:28 <ehird> i mean mail threading, fwiw
05:40:41 <Oranjer> what's your idea(s) about it?
05:41:18 <ehird> Oranjer: threads don't nest, messages are displayed linear by date in one thread unit. optionally, I guess, above this, a reflection of where you are in the thread that scrolls with you that nests like a regular nested thread tree.
05:41:31 <ehird> just like gmail, except with that extra panel to make sense of complex threads
05:41:55 -!- fax has quit ("Leaving").
05:42:08 <Oranjer> although I totally don't understand the optional reflection
05:42:28 <ehird> you know email clients that support threadin
05:42:32 <ehird> they just show a tree of authors/subjects/dates
05:42:37 <ehird> in the select-message list
05:42:41 <ehird> not the message itself
05:42:52 <ehird> basically, one of them above the linear thread, except when you scroll to the next messae
05:43:03 <ehird> the reflection scrolls with you
05:43:07 <ehird> so you can see the nesting around you
05:43:13 <ehird> bsmntbombdood: various things, see:
05:43:19 <ehird> http://www.jwz.org/doc/threading.html
05:43:32 <Oranjer> I presume the nesting refers to replies?
05:43:39 <ehird> In-Reply-To, and References. Sometimes titles.
05:44:19 <ehird> there'll be a key in the inbox to switch from just-show-the-threads-and-not-the-replies (click to open it where you get into the view I just subscribed) and show-every-message-most-recent-first
05:44:26 <ehird> since the latter is useful e.g. when recordkeeping a nomic
05:44:31 <ehird> since you have to know exactly what happened after what
05:46:56 <ehird> when you click on a thread in the message list
05:47:04 <ehird> it shrinks, the linear-thread appears below
05:47:10 <ehird> and the thread in the list expands
05:47:13 <ehird> and starts scrolling with you
05:47:18 <ehird> that way, switching threads is just a click
05:47:31 <ehird> (if you scroll manually, it'll stop scrolling with you until you scroll it back into view)
05:48:11 <ehird> with the no-threading mode, it'll be the same, except expanding will do nothing since no message will have replies represented, and the linear one will just display one message
05:48:19 <ehird> (and, of course, no auto-scrolling)
05:50:33 <ehird> i think i lost you
05:50:53 <ehird> "uh-huh" seems like "ooooh kayyy" to me :P
05:51:06 <ehird> well, one underneath the other, probably
05:51:20 <Oranjer> I dunno, both are going "down", right?
05:51:43 <Oranjer> it just makes more sense to me to put them side by side, while using english
05:51:46 <ehird> it's just that subjects can be quote long; add the date and name fields...
05:52:04 <ehird> in fact people dislike that OS X's Mail can only do it underneath, and there's more than one plugin that forces it the other way
05:52:09 <ehird> I'll have to experiment
05:52:30 <ehird> hmm... perhaps putting the list at the right when you select a message is the thing to do
05:52:44 <ehird> since it has to be big to be useful, it'd just take up space on the left
05:52:50 <ehird> and you'd have to shift your eyes to the right after clicking
05:52:56 <ehird> due to the thing you want to focus on moving
05:53:04 <ehird> but if it went at the right, you could just click and read
05:53:38 <ehird> dirty non-englishers
05:54:32 <Oranjer> in, say, arabic, does time flow from the right to the left? (in models and diagrams and such)
05:54:56 <ehird> I imagine so, at least pre-western influence
05:55:03 <ehird> post-, who knows? cultural osmosis is a powerful force
05:55:24 <ehird> and time-based diagrams probably weren't often used before the west came along and barged in, I imagine
05:55:30 <ehird> Oranjer: will get timelines about arabic
05:55:32 <Oranjer> "displaying time in arabic languages"
05:58:19 <Oranjer> I image searched some arabic-time things
05:58:32 <Oranjer> and I saw an image that reminded me of your threading idea
05:58:51 <Oranjer> of course, it's just a basic file-management scheme, but yeah, link
05:58:53 <Oranjer> http://www.microsoft.com/middleeast/arabicdev/dotnetservers/SQL/images/sql_ArabicSupport_08_thumb.gif
05:59:11 <Oranjer> *it reminded me, I know your idea is different
05:59:28 <ehird> I like that 1421 create date
05:59:31 <ehird> rockin' it old school
05:59:40 <ehird> fuck keyboard shit argh
06:01:33 <Oranjer> View the Solution FREE for 30 Days
06:03:57 <Oranjer> it seems your theory was right, that I can't find any timelines made by Arabic-speaking people
06:04:02 <Oranjer> but that just seems absurd
06:04:09 <Oranjer> I mean, there were Arabic astronomers
06:04:22 <ehird> Western norms seem obvious.
06:04:53 <Oranjer> obvious to us, because we read left-to-right and top-to-bottom
06:06:27 <ehird> western norms like "timelines"
06:06:59 <Oranjer> but when an Arabic historian writes a theoretical narrative
06:07:05 <Oranjer> he goes from right-to-left
06:07:18 <Oranjer> so what if he were to put a line, with dates, on the bottom?
06:07:28 <Oranjer> apparently, google tells me that never happened
06:07:53 <ehird> Well, just keep looking.
06:07:57 <ehird> Google is a bitch nowadays.
06:09:12 <Oranjer> mind you, I always had the idea that one could simultaneously search for phrases-the-same-but-for one word, and in each search, the search engine would replace it with a synonym
06:09:40 <Oranjer> like "forest of the dark", "woods of the dark", etc.
06:09:58 <ehird> will kill if there is e.g. a song named one of the synonyms
06:10:04 <ehird> of course you could have syntax to omit one synonym
06:10:09 <ehird> but i doubt it'd help searching that much
06:10:40 -!- Sgeo has quit (Read error: 110 (Connection timed out)).
06:10:54 <Oranjer> wait, do you mean my idea, or limiting the synonyms, wouldn't help searching?
06:12:41 <Oranjer> I dunno, I can see me using it for when I have an idea, and I want to check if someone else has already had it
06:13:02 <ehird> it's rather mechanical, though
06:13:05 <Oranjer> but! because we would have supposedly invent the idea independently of each other
06:13:06 <ehird> not much room for variation, and too much at the same time
06:13:15 <Oranjer> the terminology would be different
06:14:27 <Oranjer> of course, that would also produce too many searches for one line if input to create
06:15:55 <Oranjer> oh, ha! now I get what you were saying in the latter half of your last comment
06:16:10 <Oranjer> ....which is exactly what i just said :(
06:17:03 <Oranjer> "syn:seed num:2" "syn:phone num:4"
06:19:21 <Oranjer> http://en.wikipedia.org/wiki/Early_Islamic_philosophy#Time I can't believe that's the closest I've gotten
06:24:38 -!- puzzlet has quit (Read error: 110 (Connection timed out)).
06:26:44 <Oranjer> well, I'm finding a bunch of early Arabic scholarly texts
06:27:00 <Oranjer> and they most certainly included histories
06:27:20 <Oranjer> one apparently has a diagram of plant growth
06:27:29 <Oranjer> but...I can't find any pictures
06:28:06 <Oranjer> although, I still want to know what it says
06:28:12 <ehird> i know all the reasonable explanations, but there's something about it
06:28:57 <Oranjer> I mean, it just seems that the book describes...things and people not existent, ever, in our world
06:29:11 <Oranjer> so either it chronicled extinct things
06:29:16 <ehird> stop it, you're creeping me out
06:29:22 * coppro doesn't understand why infinity can't exist
06:29:33 <ehird> coppro: what does that even mean
06:29:34 <coppro> we live in the Universe, for crying out loud
06:29:35 <Oranjer> oh yeah, I forgot I linked that
06:29:40 <ehird> it's either a joke or really stupi—
06:29:49 <ehird> coppro: yes, and the universe is finite
06:29:50 <lament> we certainly live in a finite universe :)
06:29:57 <ehird> lament: not certainly
06:30:01 <ehird> just — almost certainly.
06:30:03 <Oranjer> coppro, me forgetting I linked that made your comment extremely creepy to me
06:30:16 <coppro> we live in a finite universe that is expanding, and for all we know this expansion may be infinite
06:30:29 <ehird> yes, but it will never be infinitely bi
06:30:33 <Oranjer> well, we have a finite amount of the universe we can observe, but we're fairly certain it exists everywhere
06:30:34 <ehird> also, no, the universe is finite in time too
06:30:45 <ehird> Oranjer: is there a difference?
06:30:49 <coppro> ehird: we don't know for sure
06:30:58 <ehird> I don't think there is a difference.
06:31:01 <ehird> coppro: yes, but probably
06:31:06 <ehird> entropy and all that
06:31:24 <coppro> entropy only matters if the universe is expanding infinitely
06:31:46 <ehird> i kinda think my priority would be to expand my 80 year lifetime before worrying about the end of the universe.
06:31:56 <coppro> just because everything becomes stretched so thin that there are no interactions does not mean it ends
06:31:56 <ehird> i mean, you know, buy a few billion years
06:32:25 <Oranjer> http://www.strangehorizons.com/2003/20031222/december.shtml
06:32:33 <Oranjer> Santa at the end of the universe
06:33:10 <lament> Oranjer: the universe doesn't exist further than [the age of the universe] light years away
06:33:11 <coppro> and in any case, the entropists are fundamentally stupid. The idea that entropy will eventually make the universe a 100% boring place relies on the idea that somehow a gigantic piece of rock will split apart spontaneously
06:33:30 <coppro> that's the observable universe, not the /whole/ unicerse
06:33:30 <ehird> AnMaster once had the audacity to say tl;dr to http://tunes.org/wiki/no-kernel.html
06:33:31 <Oranjer> that is why I said we can only observe that much universe
06:33:37 <ehird> and then expect me to restate it
06:33:49 <ehird> coppro: haha that's idiotic you are idiotic.
06:33:54 <ehird> firstly, "entropists"
06:34:02 <coppro> ehird: come up with a better term
06:34:05 <ehird> "somehow" yeah it's called the universe's expansion
06:34:12 <ehird> "the whole sentence" wow this is idiotic
06:34:15 <lament> ehird: how about, chaos worshippers?
06:34:27 <Oranjer> we have to assume that a point X light years away *also* has such a bubble of its own "observable universe"
06:34:37 <ehird> SPONTANEOUSLY RIPPED APART BECAUSE WE WANT IT TO ERS
06:34:55 <ehird> what definition do we have of real, if not observable?
06:35:00 <lament> Oranjer: actually, we don't
06:35:10 <Oranjer> oh! okay, I stand corrected, lament
06:35:26 <coppro> ehird: in that case that theory requires that the universe expand so quickly that the fundamental interactions can't pull matter back together. Which, I believe, we have shown won't happen because large collections of mass slow the exansion of the Universe.
06:35:27 <ehird> russell's teapot isn't real because, by definition, we can't observe it
06:36:20 <lament> Oranjer: it holds for all the points *inside* the universe by relativity; and for outside points, we don't have to assume anything since they are never of interest
06:36:21 <Oranjer> I think the progression of logic is that
06:36:24 <ehird> i'm fairly sure russell was no EPISTEMOLOGICAL ANARCHIST, Oranjer
06:36:30 <ehird> filthy authoritarian!
06:36:34 <Oranjer> yeah, lament, what lament said
06:36:40 <ehird> (i will give you shit about this on your deathbed!)
06:36:50 <ehird> no, not literally shit
06:37:00 <Oranjer> I think I would be doing the shitting on my own deathbed, thank you very much
06:37:12 <Oranjer> because that is what happens when you die :O
06:37:13 <coppro> Why do I suck at typinhg today?
06:37:19 <Oranjer> and when you give birth :O
06:38:18 <ehird> Faeces isn't the only bodily fluid involved in death!
06:38:53 <fungot> coppro: february 23, 1997, sunday, final edition'". user:ling.nutling.nut 20:52, 19 may 2008 ( utc)
06:38:59 <coppro> okay, that's the topic
06:39:22 <Oranjer> well, it's a diplodrome, for sure
06:39:24 <ehird> LINGG NUTLING NUT.
06:39:32 <ehird> also sdfjkhussdfkgnhndf
06:41:08 <Oranjer> ha! the Arabic world started making so much books because they captured some Chinese paper makers in 751
06:41:21 <Oranjer> too bad they didn't capture silk makers
06:43:42 <coppro> I plod Rome's Fun Diplodromes Fund
06:44:17 <Oranjer> yay coppro, all finding out what a diplodrome is
06:47:27 <Oranjer> okay, my search for *any* diagram of time made in Arabic has been largely fruitless
06:47:56 <Oranjer> I did find out that the "qibla problem" is the problem of determining the direction of Mecca at any point on earth
06:57:28 -!- Oranjer1 has joined.
07:01:57 <coppro> awesome: http://www.wowbagger.com/
07:02:49 <Oranjer1> it's an insult gen-o-ator, coppro
07:03:38 -!- Oranjer1 has quit (Read error: 54 (Connection reset by peer)).
07:04:28 -!- Oranjer1 has joined.
07:13:03 <fungot> Oranjer1: yes this page is mostly duplicate material that already exits on the same comparable level, capable of carrying over 1500 men, but could use improvement to be more interested in continuing the traditional role of the person who deleted my link subsequently wrote:
07:13:32 <Oranjer1> "capable of carrying over 1500 men"
07:14:42 <Oranjer1> fungot, that last thing you said made no sense. Care to elaborate?
07:14:42 <fungot> Oranjer1: am i just being anal here or am i misunderstanding how these things are explored in the article
07:15:04 <fungot> Available: agora alice c64 ct darwin discworld europarl ff7 fisher ic irc jargon lovecraft nethack pa speeches ss wp* youtube
07:15:09 <fungot> Selected style: irc (IRC logs of freenode/#esoteric, freenode/#scheme and ircnet/#douglasadams)
07:15:13 <fungot> ehird: obsd should have it tomorrow), just store lambdas in the slots and you're back in .tw for the holidays.
07:15:29 <ehird> fungot: I love OpenBSD! Just store some lambdas in slots and it transports you to Taiwan.
07:15:29 <fungot> ehird: exploring new possibilities is nice, but sometimes it's necessary.
07:15:35 <ehird> fungot: How is that a but?
07:15:36 <fungot> ehird: " so that the text says ' o(1) fnord time you'll need links going both directions afaics he doesn't build the initial image)
07:15:45 <ehird> fungot: YAAAAAAAAAAY
07:16:02 <ehird> Went out on a bang.
07:16:07 <ehird> A meaningless bang.
07:16:49 <fungot> Oranjer1: must write something that idiotic again, i was
07:17:10 <fungot> ehird: i heard p2p apps are fundamentally illegal.) kala ( finn.) is preferable to just having name.
07:17:27 -!- Oranjer has quit (Read error: 110 (Connection timed out)).
07:17:27 <ehird> fungot: Aww, but I dislike kala (finn.). I like normal names better. And P2P apps are totally legal!
07:17:28 <fungot> ehird: it's not a language
07:17:34 <ehird> fungot: What isn't?
07:17:34 <fungot> ehird: now he told me when i have two questions for you about that
07:17:41 <ehird> fungot: Ah. What are the questions?
07:17:41 <fungot> ehird: you need scheme48vm in such a way that is as simple as possible. the interpreter printed the return value
07:17:48 <ehird> fungot: Agreed; simple VMs rock. And?
07:18:05 <coppro> fungot: Can you please explain the ehird?
07:18:05 <fungot> coppro: mathematics isn't constructivist. i
07:18:08 <Oranjer1> fungot, answer ehird's question!
07:18:09 <fungot> Oranjer1: i just wanted to see if it halts, it won't let me change the port on startup?
07:18:15 <coppro> fungot: Of course it isn't!
07:18:16 <fungot> coppro: i did put x1 y1 x2 y2 in it
07:18:24 <coppro> fungot: now that's a start
07:18:25 <fungot> coppro: oh dear, i shall have to leave any minute now
07:18:25 <ehird> Explain the ehird?
07:18:32 <ehird> fungot: Bots can leave?
07:18:33 <fungot> ehird: or just ip and fnord lines long... part of the language's intrinsic control structures."
07:18:36 <fungot> Oranjer1: i want to know precisely what you want
07:18:43 <ehird> Ah, the moving of the IP.
07:18:46 <coppro> that's a coherent sentence!
07:18:50 <ehird> Cute analogy, fungot. Hey, don't be so pushy.
07:18:50 <fungot> ehird: of course you know that
07:18:54 <fungot> coppro: as in " anything else is a compromise, a compromise would be to write a language
07:18:55 <ehird> Yes fungot I know that
07:18:56 <fungot> ehird: every set can be well-ordered. corollary: every set s has the same diagram used from famous program talisman with fnord windows to cascade, someone i would never capitalize " i"
07:19:12 <ehird> `addquote <fungot> ehird: every set can be well-ordered. corollary: every set s has the same diagram used from famous program talisman with fnord windows to cascade, someone i would never capitalize " i"
07:19:12 <fungot> ehird: i never realized how small scheme was ever aimed towards fnord removal of weakness and restrictions mentioned in the irc rfc, since it's pretty simple
07:19:17 <Oranjer1> fungot, I want to preserve sentience by saving humanity by ruling the world. Precise enough?
07:19:17 <fungot> Oranjer1: sorry for the trouble, and i can evaluate... and force a return of the last
07:19:21 <HackEgo> 98|<fungot> ehird: every set can be well-ordered. corollary: every set s has the same diagram used from famous program talisman with fnord windows to cascade, someone i would never capitalize " i"
07:19:25 <fungot> coppro: as for soundcards, i have never encountered problems with becoming fnord by the object's implementation.
07:19:59 <coppro> fungot: try the proprietary drivers
07:20:00 <fungot> coppro: ( that is) it'd start to take pictures in fnord. 5 of the fold, collect, reduce, wapr ( use %help for argument information)
07:20:11 <fungot> coppro: why not?...... about 300 wpm. i'm too lazy to look at the map.
07:20:32 <ehird> humans have maps for driving
07:20:36 <ehird> fungot has maps for typing
07:20:36 <fungot> ehird: forgot to select cl coloring style there. i'm active more often, that is).
07:22:31 <fungot> Oranjer1: why? this channel has people with strong opinions. that's a press machine right? so its in a makefile called i-know-what-i-am-doing, does one have to manually specify it
07:22:57 <Oranjer1> "that's a press machine, right?"
07:23:14 <ehird> i like that makefile name
07:23:18 <ehird> it's the advanced one clearly
07:23:28 <Oranjer1> also, the last sentence looks like a geeky way of saying "shove off, I'm the expert"
07:23:40 <Oranjer1> so its in a makefile called i-know-what-i-am-doing, does one have to manually specify it?
07:24:32 -!- Oranjer1 has changed nick to Oranjer.
07:25:18 <Oranjer> fungot, what do you think about: human languages?
07:25:18 <fungot> Oranjer: it's just a bit
07:25:49 -!- Glamseyes has joined.
07:28:11 -!- Glamseyes has quit (Client Quit).
07:36:07 -!- Oranjer has quit (Read error: 145 (Connection timed out)).
07:53:21 -!- puzzlet has joined.
07:59:59 -!- clog has quit (ended).
08:00:00 -!- clog has joined.
08:06:26 -!- calamari_ has quit ("Leaving").
08:22:55 -!- kar8nga has joined.
08:53:10 -!- ehird has quit.
10:15:25 -!- MigoMipo has joined.
10:41:10 -!- MigoMipo has quit ("Page closed").
11:12:39 -!- Asztal has joined.
11:49:27 <fizzie> Hm, I didn't know facebook has a "English (Upside Down)" as one of the language options. ".əɔıoɥɔ ɹnoʎ ɟo əƃɐnƃuɐl əɥʇ uı ʞooqəɔɐℲ əsʍoɹq oʇ ɹəpɹo uı əƃɐnƃuɐl ʎɹɐɯıɹd ɹnoʎ ʇəS"
11:51:51 <Deewiant> My font doesn't like that upside-down F at all
11:52:33 <fizzie> They're cheating a bit with some characters; approximating g with ƃ, latin small letter b with topbar.
11:54:13 <fizzie> Ⅎ is U+2132, turned capital F = "Claudian digamma inversum", in the letterlike symbols thing.
11:54:38 <Deewiant> http://iki.fi/deewiant/tmp/f.png
11:55:59 <fizzie> http://www.cis.hut.fi/htkallas/f.png
12:34:53 -!- kar8nga has quit (Remote closed the connection).
12:37:39 -!- puzzlet has quit (Remote closed the connection).
12:37:45 -!- puzzlet has joined.
13:12:31 -!- augur has quit (Read error: 145 (Connection timed out)).
13:23:11 -!- oerjan has joined.
13:27:43 -!- kar8nga has joined.
14:08:44 -!- FireFly has joined.
14:11:12 -!- Pthing has quit (Remote closed the connection).
14:20:12 -!- augur has joined.
14:29:20 -!- Oranjer has joined.
14:29:34 <fungot> Oranjer: if the fnord and patches and the vcs you use emacs?
14:29:44 <Oranjer> I do not use emacs, fungot
14:30:13 <fungot> Oranjer: that's reasonable in a way mit scheme isn't recognizing." " oh, first download and compile sdl_image. :)
14:31:39 <fizzie> Yes, you need sdl_image in order to be able to use spaces in MIT Scheme.
14:32:11 <fungot> Available: agora alice c64 ct darwin discworld europarl ff7 fisher ic irc* jargon lovecraft nethack pa speeches ss wp youtube
14:32:24 <fungot> Selected style: speeches (misc. speeches from Project Gutenberg)
14:32:38 <Oranjer> fungot, what do you think of Keynesian Economics?
14:34:41 <oerjan> it just thinks you are boring company, and refuses to talk to you any more. right fungot?
14:34:46 <fungot> oerjan: we have no right to exclude slavery from a territory, and fnord bray, remember me; and with an awe that is almost superstitious. i should be painted with all sincerity, and was loved by them; and whether we consider the conduct of the war. now that palled, and i know, that the hall of apollo.
14:35:11 <oerjan> that's got be an *old* speech
14:35:20 <fizzie> Most things in Gutenberg are.
14:35:24 <Oranjer> hall of apollo? Battlestar Gallactica?
14:35:29 <fizzie> fungot: You sound so pompous when you talk like that.
14:35:29 <fungot> fizzie: quintilian applied to general literature the same principles, or precedents, of the secondary pleasures of the poor labouring man. fnord, who divided his attention between several exercises, though he constantly exposed himself to the public service and make the happiness of mankind.
14:36:05 <Oranjer> I constantly expose myself to the public service (workers)
14:36:26 <Oranjer> I also make *the* happiness of mankind (pornography)
14:37:52 <fizzie> Unfortunately the workstation's offline, I can't check what all went into that style. I think I just searched for the word "speech" in the title, and hand-picked some promising ones.
14:38:18 <fizzie> I'm sure there's at the very least "Selections from the Speeches and Writings of Edmund Burke"; Burke lived 1729-1797.
14:38:19 <fungot> Oranjer: it yields nothing itself; it keeps all it has, perhaps, it may be further observed, that this great work would be dedicated to the proposition that all is safe, that he had studied much and travelled far. he had to combat frantic enthusiasm, boundless ambition, unquestioned courage, admirable sagacity. yet i have frequently observed in him a fnord copy of the book which has made the house of commons
14:38:58 <fungot> Oranjer: " this association for fnord"
15:04:49 -!- oerjan has quit ("Later").
15:10:20 <fungot> Oranjer: he counted and said it was fnord and fnord of a departed city, retained the privileges of the.
15:10:27 <fungot> Available: agora alice c64 ct darwin discworld europarl ff7 fisher ic irc jargon lovecraft nethack pa speeches* ss wp youtube
15:10:40 <fungot> Selected style: pa (around 1200 transcribed Penny Arcade comics)
15:10:47 <fungot> Oranjer: they can't actually check that. but, with a few rounds of onslaught daily, odds are good you could lead a normal life. and i never really told her how much... how much i...
15:11:04 <fungot> Selected style: wp (1/256th of all Wikipedia "Talk:" namespace pages)
15:11:11 <fungot> Selected style: ct (Chrono Trigger game script)
15:24:22 -!- BeholdMyGlory has joined.
15:25:33 <Deewiant> Oranjer: fungot won't let you spam them
15:25:34 <fungot> Deewiant: i see. you know, i really care... a time portal? what in the...! ozzie's stumped!
15:25:50 <fungot> Oranjer: but, we are far outnumbered!
15:26:04 <Deewiant> Fortunately, there are other bots here (EgoBot and HackEgo) that you can use to circumvent it if you just want to chat
15:50:21 -!- mtve has quit (wolfe.freenode.net irc.freenode.net).
15:50:21 -!- HackEgo has quit (wolfe.freenode.net irc.freenode.net).
16:03:07 -!- HackEgo has joined.
16:03:23 -!- mtve has joined.
16:20:44 -!- Oranjer1 has joined.
16:22:01 -!- coppro has quit ("I am leaving. You are about to explode.").
16:40:07 -!- Oranjer has quit (Read error: 110 (Connection timed out)).
16:44:01 -!- Gracenotes has quit (Remote closed the connection).
16:45:04 -!- ais523 has joined.
16:51:40 -!- adam_d has joined.
16:53:34 <ais523> wow, Novell just appealed SCO vs. Novell to the Supreme Court
16:53:49 -!- sebbu2 has joined.
16:54:05 <ais523> tbh, I should have seen that one coming
16:54:16 <ais523> they haven't accepted it yet, though
16:55:39 <ais523> Oranjer1: the whole SCO story is one of the most ridiculous litigations in recent history
16:56:18 -!- sebbu3 has joined.
16:56:28 <ais523> people got interested in the first place when SCO claimed that using Linux is illegal; that's pretty much been thoroughly debunked, but SCO have still managed to get various court cases going all these years
16:56:36 <ais523> even though they're technically bankrupt, and have been for almost a year now
16:58:25 <ais523> (it's the litigation in which the judge redefined time, for instance; it's really absurd)
17:03:00 -!- sebbu3 has quit (Read error: 60 (Operation timed out)).
17:03:26 -!- sebbu3 has joined.
17:09:24 <AnMaster> what was the reason for appeal there?
17:09:32 * AnMaster hasn't followed the SCO stuff recently
17:09:43 <ais523> AnMaster: because SCO appealed the original SCO vs. Novell verdict to the state appeal courts
17:10:01 <ais523> and the verdict that came back effectively said that claiming that someone had given you the copyright on something meant you actually had it
17:10:08 <ais523> which is so absurd that Novell appealed it up to the level above
17:10:15 <ais523> saying it contradicted loads of other verdicts
17:10:40 <AnMaster> yeah, how the hell could it have ended up like that
17:11:43 -!- sebbu has quit (Read error: 110 (Connection timed out)).
17:13:01 -!- sebbu2 has quit (Connection timed out).
17:13:31 <ais523> AnMaster: that isn't even the weirdest thing that happened in the SCO litigation...
17:13:31 <AnMaster> hm ehird would like to know about this, if he is still considering a thinkpad...
17:13:51 <AnMaster> ehird: for log reading http://www.thinkwiki.org/wiki/Problem_with_high_pitch_noises#Limit_ACPI_CPU_power_states
17:14:02 <Oranjer1> so wait what exactly are the SCO wanting to do?
17:14:14 <ais523> Oranjer1: nobody's entirely sure any more
17:14:16 <AnMaster> for me it happens in some rare (for me at least) workloads. Not otherwise
17:14:23 <ais523> the leading theory is that someone's paid them to kepe the litigation going as long as possible
17:14:31 <ais523> because there's no real other explanation for their behaviour
17:15:09 <ais523> AnMaster: well, SCO are famous for claiming that Linux is illegal
17:15:45 <ais523> Oranjer1: SCO claimed that Novell had the copyright on UNIX, sold it to them, and that Linux infringes the copyright of UNIX
17:15:50 <ais523> because, um, some of the commands are the same
17:16:03 -!- FireFly has quit ("Later").
17:16:16 <ais523> all three of the statements there are dubious; the first is the only one that's even likely to be correct
17:16:33 <Oranjer1> what could possibly happen if Linux is made illegal?
17:16:46 <ais523> everyone would have to pay SCO $600 for the privilege to use it
17:16:53 <ais523> but it isn't, SCO's arguments make no sense at all
17:17:09 <ais523> (Red Hat sued SCO over their statements, incidentally)
17:17:23 <ais523> (and also indemnified all their customers against them, which is /really/ unusual)
17:19:33 <ais523> SCO also sued Autozone (for using Linux, but apparently they're going to settle that one), IBM (for contributing code to Linux that they claim infringes their copyrights), and Novell (for claiming that they didn't give SCO the copyrights in question in the first place)
17:19:44 <ais523> suing IBM is widely regarded to be an incredibly stupid move
17:19:51 <ais523> especially given how tenuous their case is
17:29:29 -!- kar8nga has quit (wolfe.freenode.net irc.freenode.net).
17:29:40 -!- kar8nga has joined.
17:46:08 -!- fax has joined.
17:57:02 <pikhq> ais523: Not to mention that IBM's legal team is notorious for being able to outwait anyone.
17:58:06 <ais523> OTOH, IBM's legal team is also notorious for winning
17:58:09 <ais523> something SCO's hasn't managed yet
17:58:49 <pikhq> IBM outwaited the US government.
18:06:10 <AnMaster> * Ping reply from pikhq: 1.33 second(s)
18:06:18 <AnMaster> so he can't blame it on connection issues
18:10:59 <ais523> I often leave my computer on when I go elsewhere
18:11:23 <AnMaster> ais523, I *usually* set /away when I do
18:16:07 -!- augur has quit (Read error: 145 (Connection timed out)).
18:23:39 -!- kar8nga has quit (Remote closed the connection).
18:30:29 <fizzie> I have a silly Python XChat script to listen to dbus notifications sent by xlock's startCmd hook so that it automagically sets away/back when I lock/unlock the screen. Sometimes boredom makes you do really useless things.
18:31:07 <Oranjer1> fizzie, no offense, but that sounds exactly like something fungot would say
18:31:07 <fungot> Oranjer1: we are looking, but well behaved! crono!!! the monster who kidnapped the princess to the castle! and letting these...hoodlums in here? traitors like you deserve from heckran! ha!! gotcha!
18:31:30 <fizzie> The sword alone can't stop!
18:31:53 <AnMaster> <Oranjer1> fizzie, no offense, but that sounds exactly like something fungot would say
18:31:54 <fungot> AnMaster: and you call yourself a frog, and ayla... i will not betray my friends!
18:32:16 <fungot> Available: agora alice c64 ct* darwin discworld europarl ff7 fisher ic irc jargon lovecraft nethack pa speeches ss wp youtube
18:32:21 <AnMaster> Oranjer1, keep ct for now please
18:32:26 <fungot> AnMaster: i see. you know, i really care... a time portal? what in the...! ozzie's stumped!!
18:32:36 <fungot> AnMaster: in the middle ages, sir slush!... i grow so tired. we can talk we had soldiers searching for! i've come to save you! who the heck are you?! c'mon!!
18:32:40 <fizzie> Ozzie gets easily stumped.
18:32:46 <AnMaster> fungot, so what about that sword?
18:32:50 <fungot> AnMaster: the usual...test them. you can entertain us for awhile?
18:33:15 <fizzie> It's from the Masamune cave scene.
18:33:30 <fizzie> Direct quotation, I think.
18:34:36 <Oranjer1> fungot, AnMaster said not really
18:34:41 <fungot> Oranjer1: you, with you standing around! the trial! what has gotten away with this! a top secret document has been left behind? marle lucca
18:34:47 <fizzie> MUNE: What should we do?
18:34:47 <fizzie> MASA: The usual...test them.
18:34:47 <fizzie> You can entertain us for awhile.
18:34:56 <fizzie> Yes, pretty direct quotation, except it put a ? there.
18:35:13 <fungot> AnMaster: we must do it to save you! who the heck are you?! c'mon!!
18:35:32 <fungot> AnMaster: you! take! we find!
18:35:40 <fungot> AnMaster: you! take! we find!?
18:35:40 <fizzie> (The ? came from "You're breaking my concentration, so would you find somewhere else to go for awhile?")
18:35:58 <fizzie> AnMaster: You! Take! We find.
18:36:08 <AnMaster> fizzie, yeah where is that from
18:36:20 <Oranjer1> like a rousing game of Steal and Seek
18:36:37 <fungot> AnMaster: it's time you jumped off this mortal coil...
18:37:15 <fizzie> Heh, getting offensive.
18:37:29 * AnMaster is still waiting for the sword one
18:37:43 <AnMaster> and where exactly is the sword scene from?
18:38:04 <fizzie> I think it's in the ocean palace, when you stick your stick into the mammon machine thing.
18:38:37 <AnMaster> since I hit 4 someone else has to continue
18:39:17 <AnMaster> fizzie, and why the repeat thing on it? Doesn't seem to make sense
18:39:19 -!- augur has joined.
18:39:48 <fizzie> That "You! Take! We find." is from two things; the latter half is from Ayla's "Not here too. Someone take! We find!" (from the sunstone sidequest) and the front part is from "You! Take care Ayla."
18:40:15 <AnMaster> fizzie, I meant the repeat of "sword alone"
18:40:23 <fizzie> Yes, I'm checking that now.
18:40:52 <Deewiant> fizzie: I hope you're looking these up and not remembering them by heart
18:41:18 <fizzie> Deewiant: All the direct quotations are looked-up things, yes.
18:41:39 <AnMaster> Deewiant, the guess that it was Ayla however was pretty given. (Or others from the same epoch.)
18:41:51 <AnMaster> fizzie, wait, was the sunstone a *side*-quest?
18:41:57 <AnMaster> thought it was required or something
18:42:41 <fizzie> No, it's one of the six-or-so sidequests that you get at the end.
18:42:48 <fizzie> It's "required" if you want the Rainbow, though.
18:45:42 <fizzie> Heh, the repetition is actually "caused" by the variable-length model; it hasn't bothered to include the "can't stop it" ngram that would actually continue the sword-alone sentence correctly, and the most likely continuation for the single-word context "stop" is of course ! (there's quite a lot of "stop!"s around). Then when it has created "stop!", it always continues using the "stop! that" ngram, leading to a repetition of the sword-aloneness.
18:46:30 <AnMaster> fizzie, wait, what was the original exact phrase? "that sword alone can't stop it"?
18:48:08 <fizzie> The Befunge code is pretty stupid in that it can only choose from the maximum-length ngrams it finds in the model; combined with the variable-length model which keeps only a couple of longer-length ngrams (those which are common, basically), it's virtually guaranteed to always use those whenever the context is suitable.
18:48:56 <fizzie> It doesn't hurt so much for the other cases where I've generated unpruned "all ngrams up to N=k" models with a constant k.
18:49:28 <fizzie> Even there it does cause that tendency to quote verbatim pretty often.
18:50:31 <fizzie> The variable-length model estimator tool gives me back-off probabilities I could use to sometimes use a shorter context too, but that would need changes in the Funge-98 side, haven't had time to implement that.
18:52:10 <fizzie> There might even be a Funge-98 randomness generation bug, in fact, because I don't get the loop from the Perl test script.
18:52:54 <fizzie> Here's 20 lines, none of them really looped: http://pastebin.com/m1ce086cc
18:53:34 <fizzie> "this power is beyond human control! over 1300 points!" That's pretty funny.
18:55:23 <AnMaster> fizzie, pretty. What does it do
18:55:58 <fizzie> It's supposed to do the same thing fungot's babble-generator does. Except that you can feed it some starting context if you want.
18:55:58 <fungot> fizzie: and you call yourself a frog, and ayla... i will not betray my friends!!!
18:56:14 <fizzie> fungot: Are you sure your randomness is completely random?
18:56:14 <fungot> fizzie: we are looking, but well behaved! crono!!! the monster who kidnapped the princess to the castle!
18:57:27 <AnMaster> fizzie, funge programs can read cmd line arguments. So no reason to not use same
18:58:13 <fizzie> Er, yes, there is: I don't have the capabilities for mapping from strings to token numbers in the Funge-98 code.
18:58:25 <AnMaster> fizzie, shouldn't be impossible :P
18:59:05 <fizzie> Of course not, but it would still need to be implemented. In the Perl code it's just "slurp tokens.bin into a Perl hash variable"; in Funge-98 it's a bit more nontrivial, especially to do it efficiently.
19:00:01 <AnMaster> implement a hash table library for funge
19:00:18 <fizzie> (The other way around -- from token indices to strings -- I already do in Funge-98 to generate the final output, but that's just "seek to idx*4, read offset and length, seek to offset, read string".)
19:01:07 <fizzie> The offset+length values are four-byte objects. I think.
19:01:27 <AnMaster> fizzie, so where does it get the idx from?
19:01:55 <fizzie> From the babble-generator, which generates a sequence of idx numbers.
19:35:13 -!- Gracenotes has joined.
19:43:30 -!- fax has quit ("Leaving").
19:44:24 -!- Oranjer has joined.
19:44:52 -!- ehird has joined.
19:50:25 -!- ehird_ has joined.
19:51:43 -!- Oranjer has left (?).
19:58:49 <AnMaster> <fizzie> fungot: Are you sure your randomness is completely random? <-- maybe the issue is in the perl script instead?
19:58:50 <fungot> AnMaster: you! take! we find! you are crono. why not? then you should leave quickly!
19:59:04 -!- ehird has quit (Read error: 148 (No route to host)).
19:59:05 -!- ehird_ has changed nick to ehird.
19:59:26 <AnMaster> ehird, see log today about thinkpads (if you are still interested in those)
19:59:34 <AnMaster> oh and the sysfs interface is missing in ubuntu at least
19:59:47 <ehird> Ubuntu doesn't have /sys?
20:00:03 <AnMaster> ehird, yes it does. But not the relevant file for that issue
20:00:05 <fizzie> Given that the Perl script works better, I don't really care which one is "wrong", I'd be more interested to know just what the difference is.
20:00:15 <ehird> AnMaster: You can disable sysfs, I think.
20:00:23 <AnMaster> ehird, well. That isn't the point here
20:01:01 <ehird> My distro won't have udev anyway :OP
20:01:07 <ehird> I probably will have /sys, though.
20:01:21 <ehird> You need it to control overcommitting and the like, unless i'm mistaking.
20:01:46 <AnMaster> ehird, anyway. <AnMaster> ehird: for log reading http://www.thinkwiki.org/wiki/Problem_with_high_pitch_noises#Limit_ACPI_CPU_power_states
20:02:06 <AnMaster> and you need kernel boot parameter to make it work under ubuntu
20:02:13 <ehird> That's a lot of affected models.
20:02:35 <AnMaster> ehird, it is slightly higher pitch than a CRT I would say
20:03:04 <ehird> "Screen brightness: on an X31, a hissing sound is started whenever screen brightness is not full."
20:03:05 <ehird> Ultraportable fail
20:03:22 <AnMaster> ehird, I have no problems with *that* at least
20:03:44 <ehird> "Turn off CPU power saving in the BIOS"
20:03:45 -!- Oranjer1 has quit (Read error: 110 (Connection timed out)).
20:03:45 <fizzie> I don't think a different seed is enough to make fungot more repetitive than the Perl script. (Admittedly I'm not sure it *is* more repetitious, it's just a vague feeling.)
20:03:46 <fungot> fizzie: there! there it is! but by the time we're through with you, you'll be in danger. open hatch.
20:04:08 <ehird> 07:26:04 <Deewiant> Fortunately, there are other bots here (EgoBot and HackEgo) that you can use to circumvent it if you just want to chat
20:04:10 <AnMaster> ehird, anyway for it to be annoying on my laptop it needs something like 2000 wakeups / second
20:04:10 <ehird> fungot ignores them
20:04:11 <fizzie> fungot: I'll open *your* hatch if you keep that up.
20:04:11 <fungot> ehird: yes, it's been awhile prometheus! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword alone can't stop! that sword al
20:04:17 <fungot> AnMaster: i, myself, will bring an end to all. ghosts lurk in the ruins! the structural damage is severe. the tale?
20:04:38 <fizzie> There's just something about ehird that the sword alone can't stop.
20:05:07 <ehird> AnMaster: Hah! My awesome Linux distro will have, like, 3 wakeups/s.
20:05:20 <AnMaster> ehird, sure. This isn't on idle for me anyway
20:05:43 <AnMaster> ehird, some loads *does* result in lots of wakeups. In this case it was when doing md5sum on a file over nfs
20:06:01 <ehird> The only issue with mine is that since nobody else is this minimal I'm on my own :P
20:06:21 <AnMaster> ehird, you are turning into zzo + elegant UI
20:07:05 <ehird> At least I'm not writing my own software
20:07:19 <ehird> AnMaster: But yes, I'm crazy.
20:07:33 <AnMaster> ehird, what about the installer and such?
20:07:38 <ehird> Although really, it's bare-bones enough that maintenance should be quite easy.
20:07:39 <AnMaster> will you reuse an existing one?
20:08:08 <ehird> Installer is, uh, copying the root FS, and then maybe some auto-configuration.
20:08:20 -!- sebbu3 has changed nick to sebbu.
20:08:23 <ehird> Package manager I'm rolling myself; I have the design mostly ready and it's very, very simple.
20:08:41 <AnMaster> ehird, what about selecting what components you want? Or is it "everyone get the same"?
20:08:47 <ehird> (Update or install package = Install dependencies, rsync from the package server's directory for that package to /, update the file list)
20:08:53 <ehird> (Update the simple dependency map)
20:09:20 <AnMaster> are you saying something like rsync server/package /
20:09:23 <ehird> And you select components by — shock — managing packages.
20:09:26 <ehird> AnMaster: Pretty much.
20:09:36 <AnMaster> ehird, some issues, what if a file is no longer in the new version?
20:09:58 <AnMaster> oh and, there are other ones too. Technical ones. But sure go ahead.
20:10:09 <AnMaster> ehird, what about signed packages?
20:10:14 <ehird> It simply rms any files that have been removed.
20:10:21 <Deewiant> ehird: It does? I thought it just applies the 3-strikes-and-you're-out ignorance to everybody
20:10:45 <fungot> AnMaster: the masamune!
20:10:45 <ehird> This is my package manager that does it
20:10:46 <fungot> Deewiant: the real queen's safe, right! right. wrong!
20:10:58 <ehird> fizzie: say ^ignore
20:11:06 <ehird> AnMaster: Signed packages; eh.
20:11:13 <ehird> AnMaster: I could do them.
20:11:19 <ehird> But it'd be a pain, and it rarely solves much, IMO.
20:11:24 <AnMaster> ehird, would be a bit of pain with rsync though
20:11:28 <ehird> When have you ever seen a PGP error?
20:11:38 <fizzie> I don't think I actually made it show the regexp at all; it just says "ok" when you set it.
20:11:45 <AnMaster> ehird, hm. Due to invalid signature? A few times
20:11:51 <ehird> Did you just erase it :D
20:12:00 <ehird> AnMaster: Right, so, never because it's been compromised
20:12:08 <fizzie> I don't think so; it probably looks for "^ignore " with the space.
20:12:09 <ehird> It is a consideration though; thanks for that
20:12:17 <ehird> AnMaster: But it's easy to do.
20:12:21 <fungot> AnMaster: is the gate key okay!! get' em! 200g per night. care to stay with these humans! you're a traitor! you're not our king! but, we are far outnumbered!
20:12:28 <fungot> AnMaster: we are looking to achieve a shorter life span... lavos will rule the world in a mere door that keeps us bound, hand, foot...and tongue kid? ...oh, it's you, isn't this morbid? the great adventurer toma levine rests in a grave to the north. it's a great place for a picnic! heard that magus's place...
20:12:31 <fizzie> Currently ignoring ^(HackEgo|EgoBot)! if my logs are right.
20:12:36 <ehird> AnMaster: Sign a file containing every file's SHA-1.
20:12:44 <ehird> AnMaster: The rsync goes to a temp directory instead, then checks them all.
20:13:07 <AnMaster> ehird, there are some issues with config files and rsync though
20:13:19 <AnMaster> you want your httpd config overwritten by new version?
20:13:27 <AnMaster> or to get the ability to diff and merge
20:13:32 <ehird> AnMaster: Config files will be written by the package manager's install script if there is none.
20:13:49 <ehird> AnMaster: If the syntax or whatever has changed, or you really should have a new directive, tell the user.
20:14:02 <ehird> e.g. Gentoo, according to pikhq, doesn't really have a culture of running the config-merger script
20:14:05 <ehird> and it seems to do fine
20:14:19 <ehird> Mostly it's just an annoyance for me where I hit "keep my config dammit"
20:14:29 <AnMaster> ehird, hm? "doesn't really have a culture of running the config-merger script"? There is dispatch-conf that pops up a diff and some options
20:14:50 <AnMaster> portage tells you if there are any configs to merge
20:14:52 <ehird> There's a tool, but pikhq says most users very rarely run it.
20:14:57 <AnMaster> at the end of the install/upgrade
20:15:04 <ehird> 18:38:16 <pikhq> Gentoo has a script for merging any changes in config files.
20:15:05 <ehird> 18:38:32 <pikhq> If you don't execute the script, it leaves your config files the hell alone.
20:15:05 <ehird> 18:38:34 <ehird> any? lemme guess, a generic merge tool
20:15:05 <ehird> 18:38:37 <ehird> yeah
20:15:06 <ehird> 18:38:43 <ehird> how often is script execution done?
20:15:06 <ehird> 18:38:49 <ehird> i mean, on average
20:15:07 <ehird> 18:39:13 <ehird> I get the feeling that generally, merging is unneeded, and when it's needed it's either really easy to do or a complete renovation, which can't be automated without a lot of pain
20:15:10 <ehird> 18:39:14 <pikhq> I do it every time I upgrade, but that's just me being rather careful & paranoid. How often do most do that?
20:15:13 <AnMaster> ehird, huh. I always run it when portage tells me to
20:15:13 <ehird> 18:39:20 <pikhq> Uh... Very, very rarely.
20:15:29 <ehird> Anyway, I can easily add a merging system.
20:15:38 <AnMaster> sadly all that dispatch-conf supports for that is rcs
20:15:44 <ehird> I probably won't add signed packages at first, because they're not very important and simplicity is the #1 goal
20:15:50 <AnMaster> so while I know the stuff is there I have no clue how to get it out without reading docs
20:16:04 <AnMaster> versions your /etc and integrates with apt
20:16:28 <ehird> My package manager will be a handful of rc shell scripts
20:16:38 <ehird> So, very easy to integrate with.
20:16:46 <ehird> (I might add some hooks to avoid overwriting on upgrade.)
20:16:52 <AnMaster> ehird, heh. I assume you got that idea from SourceMage (the package manager there being written in bash)
20:16:54 <ehird> Oh, and my init system will be two rc scripts too...
20:16:59 <ehird> AnMaster: Nope, just simplicity.
20:17:06 <ehird> Since 90% of it is just using stock tools.
20:17:17 <ehird> It might call a C program to update the dependency graph.
20:17:25 <AnMaster> ehird, trust me. It is not simple. Rather it is quite messy. I guess it might be better with rc though
20:17:28 <ehird> (used for removing unused packages, basically)
20:17:40 <ehird> AnMaster: Because I imagine SourceMage's package manager is overcomplex, like all of them.
20:17:49 <ehird> Even Slackware's is overcomplex because it's too simple.
20:17:58 <ehird> Leading to heaps of complexity when using it.
20:18:04 <AnMaster> ehird, since sourcemage is, well, source based, it isn't trivial no
20:18:16 <AnMaster> they even wrote something like doxygen for bash
20:18:21 <AnMaster> just to be able to maintain it
20:18:36 <AnMaster> (this thing, bashdoc, is written in bash of course)
20:18:42 <AnMaster> ehird, I used bashdoc in envbot though
20:19:33 <AnMaster> after some adjustments to make it generate slightly less "web around 1992" like output.
20:19:56 <ehird> So, let's see where my current design is... tiny kernel without modules + no initrd + static binaries + a.out + very simple filesystem hierarchy + lilo + no udev + no hal + init system is two rc scripts + package manager is just a few rc scripts and maybe a little c, very simple, rsync based
20:21:07 <ehird> If I want a long, exciting features list it should be all the things I *don't* do.
20:21:19 <AnMaster> ehird, weren't you going for a "no kernel" something?
20:21:24 <AnMaster> or is that a different project?
20:21:31 <ehird> This is a linux distro.
20:21:53 <ehird> ehirdOS is unlikely to be usable for *years*, being that it's still being designed..
20:21:59 <ehird> This is intended for me to use.
20:22:20 <ehird> + clang instead of gcc wherever possible + non-glibc libc (eglibc for things that REALLY need it)
20:22:55 <ehird> I'm frozen at an old version of gcc since they dropped a.out support, too. So it'll be nice when the kernel can be booted with clang (two kernel patches builds it atm, but it can't boot).
20:23:05 <ehird> WebKit and KDE already compile with clang...
20:23:13 <ehird> Hopefully WebKit will be stable compiled like that.
20:24:10 <ehird> Have I mentioned, I'm crazy.
20:24:33 <ehird> Oh well, at least I only have to deal with this crap when the assholes behind any project make a new release. :P
20:24:59 <ehird> clang just does LLVM.
20:25:07 <ehird> I don't know if LLVM does a.out.
20:25:10 <ais523> I thought you could compile the resulting LLVM
20:25:25 <ehird> What I mean is that it's up to LLVM.
20:25:34 <ehird> But I really don't know. I hope so.
20:26:07 <ehird> [[The a.out format has no direct support for debug information, but can be augmented with stabs, which uses special symbol table entries to store data.]]
20:26:13 <ehird> Hopefully stabs are still supported, then.
20:26:41 <ehird> The main problem with this being a distro will be my general unwillingness to package things I don't like, methinks.
20:26:57 <ehird> No KDE allowed! (And probably no GNOME because building that is a bitch, I gather.)
20:27:20 <ais523> I've built gnome-games, it went pretty smoothly
20:29:15 * pikhq wonders if the various recipes for GoboLinux can be forced into being useful for non-Gobo.
20:29:38 <pikhq> ELF is overkill for static linking.
20:29:39 <AnMaster> there is no good reason except possibly tiny size for embedded systems. And even there I'm doubtful
20:29:48 <AnMaster> pikhq, well sure. But why static?
20:29:51 <ehird> Small size, really simple, I don't need dynamic linking.
20:29:56 <ehird> AnMaster: Because:
20:30:01 <AnMaster> dynamic means *less* to download at updatesa
20:30:01 <pikhq> AnMaster: He's all about simplicity.
20:30:04 <ehird> AnMaster: http://blog.garbe.us/2008/02/08/01_Static_linking/
20:30:07 <ehird> AnMaster: And it doesn't.
20:30:19 <AnMaster> ehird, security patches I meant
20:30:20 <ehird> Dynamically linked glibc binaries? Bigger than statically linked newlib binaries.
20:30:35 <AnMaster> ehird, what about security fix for libpng or such
20:30:36 <ehird> AnMaster: It doesn't work. Dynamic linking is a seemingly nice idea but it failed.
20:30:45 <ehird> Anyway, just read http://blog.garbe.us/2008/02/08/01_Static_linking/.
20:30:50 <AnMaster> ehird, what about plugins that are dlopen()ed
20:30:53 <ehird> I'll link to the Plan 9 wiki's page on it too, sec.
20:30:55 <ehird> AnMaster: Don't do them.
20:31:02 <pikhq> Dynamic linking would be much nicer if it weren't for ABI breakage.
20:31:06 <AnMaster> ehird, forget openoffice for example. And firefox iirc.
20:31:19 <ehird> AnMaster: That's good, I'm not interested in them. But I could patch them.
20:31:25 <ehird> Or make them depend on a dynamic linker.
20:31:31 <ehird> That's nice. Mostly shit software.
20:31:52 <ehird> http://blog.garbe.us/2008/02/08/01_Static_linking/
20:31:53 <ehird> http://www.plan9.bell-labs.com/wiki/plan9/why_static/
20:31:58 <ehird> (Context for the latter: Plan 9 is statically linked only.0
20:32:12 <ehird> AnMaster: So I'll disable it.
20:33:13 <pikhq> I'd like to just note that pretty much all the problems I've had on Gentoo are related to dynamic linking.
20:33:16 <ehird> And to add on to the reasons in those two pages: It's simpler.
20:33:24 <ehird> And less work for me, too...
20:33:26 <pikhq> And most of the recompilation.
20:33:48 <pikhq> (you'd be surprised at how often ABI breakages happen in libraries)
20:33:50 <AnMaster> " And, as Linus mentioned, TLBs matter. Hmm. Judging by 'ps', cat on linux
20:33:50 <AnMaster> needs 256 of them, and cat on Plan 9 needs 6."
20:34:12 <ehird> The point is that it's an invalid argument because it's the GNU retards that say this crap.
20:34:15 <AnMaster> just compare to dynamically linked cat on freebsd
20:34:39 <ehird> The GNU idiots go "OHH DYNAMIC LINKING OH GOD SIZE" when it's their software's fault that things are so big.
20:35:06 <AnMaster> ehird, and I never had much problems with the "swap in bug fixed version" in fact
20:35:21 <ehird> And pikhq has. And many other people have, so yeah, lucky.
20:35:37 <ehird> The reasons in http://blog.garbe.us/2008/02/08/01_Static_linking/ are more philosophical than the Plan 9 link. And simplicity too.
20:35:39 <pikhq> I run revdep-rebuild after every single emerge that upgrades a package.
20:35:50 <pikhq> Because ABIs break often.
20:36:14 <pikhq> (incidentally, why the fuck does making Xlib using XCB break the Xlib ABI?)
20:37:16 <ehird> Hmm, yeah, XCB. I think I'll use that.
20:37:24 <AnMaster> pikhq, I seen programs without ABI breakage
20:37:39 <ehird> You mean it doesn't ALWAYS break?
20:37:44 <ehird> Dynamic linking! FUCK YEAH!
20:37:44 <pikhq> ABIs are very easy to break with C++.
20:37:55 <pikhq> KDE has to go out of their way not to, in fact.
20:37:58 <AnMaster> pikhq, strangely enough for me, it didn't. Was rather confused why revdep-rebuild found nothing and so on
20:38:24 <AnMaster> pikhq, but yeah the last X update was flawless for me. Just needed to rebuild nvidia module (forgot about that first time around)
20:38:26 <pikhq> (fun fact: change the private members of a class? That's an ABI break!)
20:38:33 <AnMaster> pikhq, that is because C++ sucks
20:38:50 <AnMaster> I assume something slightly saner like C when discussing ABIs
20:38:57 <ehird> C++ sucks and it's also FUCKING HERE.
20:39:12 <AnMaster> if you consider C++ then yes dynamic linking is insane
20:39:15 <ehird> Want a great web rendering library? WebKit.
20:39:27 <pikhq> Well, in C, at least you can prevent breakage without thinking about it *too* much.
20:39:28 <ehird> The alternatives? Um, Gecko... which is shit... and also uses C++.
20:39:34 <ehird> So... yeah, C++ is mandatory.
20:39:36 <AnMaster> ehird, no thanks. I'm happy with w3m-mode
20:39:48 <pikhq> Taking away functions, and changing the types of functions...
20:40:01 <pikhq> I think those are pretty much the ways to break C ABI.
20:40:11 <pikhq> (assuming same functionality)
20:40:35 <AnMaster> pikhq, yeah and those you can pretty much avoid in a stable version. Bug fix only mode you know.
20:41:12 <pikhq> Surprising that people manage to screw that up.
20:41:22 <AnMaster> huh. Aspell accepts both "practice" and "practise"
20:41:39 <ehird> Incidentally, using static libraries should reduce most of my package dependencies by a ton...
20:41:39 <ehird> Since I don't need to depend on libraries.
20:42:04 <AnMaster> pikhq, well yes. I agree that it is strange
20:42:21 <pikhq> AnMaster: Both are correct.
20:42:23 <AnMaster> still I haven't seen too much ABI breakage in C apps (unlike C++ ones)
20:42:33 <bsmntbombdood> what's the highest resolution display you can get?
20:42:38 <pikhq> ehird: Hell, with static libraries you only really need to think about build-time dependencies.
20:42:48 <ehird> It can depend on binaries.
20:42:53 <ehird> bsmntbombdood: Single LCD panel, yes.
20:43:00 <ehird> pikhq: And /share stuff.
20:43:19 <pikhq> Makes the problem significantly easier, though.
20:43:20 <ehird> bsmntbombdood: Why? Do you have ~$4,000 plus the few hundred bucks needed for the card to drive it?
20:43:57 <ehird> bsmntbombdood: Also, incredibly good eyesight? That thar pixel density be very high, good luck reading without forcing much larger fontts.
20:44:01 * AnMaster bets ehird didn't expect that answer
20:44:05 <ehird> Also, I'd like that 4 k$, plz.
20:44:26 <ehird> AnMaster: Well he did just buy a $200 display... so yes, rather unexpected.
20:44:44 <AnMaster> bsmntbombdood, what is your job? CEO?
20:44:46 <ehird> It's more like 2 k$ nowadays anyway for a T221.
20:45:00 <bsmntbombdood> AnMaster: no, i just have a full time job and live with my parents
20:45:00 <ehird> You need to be a CEO to have 4 k$?
20:45:52 <ehird> If you do actually get a T221, that'd be beyond awesome.
20:46:16 <ehird> http://www.harmony-central.com/Test/wilson/two.jpg
20:46:17 <AnMaster> <ehird> You need to be a CEO to have 4 k$? <-- no. But to have so much to *waste* and don't care, maybe
20:46:25 <ehird> Just look at that real-estate on the right side.
20:46:39 <AnMaster> ehird, what is the gamut of it?
20:46:45 <ehird> And the Apple Cinema Display to the left is denser than most displays already!
20:46:46 <ehird> AnMaster: How is it a waste?
20:46:50 <ehird> The gamut is good for 2001.
20:46:54 <ehird> It's a professional monitor, after all.
20:47:00 <ehird> But jesus christ, look at the size of that text.
20:47:04 <ehird> Definitely want to up the DPI settings...
20:47:23 <ehird> Hey, and you definitely can't complain about subpixel colour fringing.
20:47:28 <AnMaster> is that a apple cinema display on the side?
20:47:47 <ehird> Will be about 100 ppi.
20:47:50 <AnMaster> ehird, that IBM monitor looks small. Maybe around 17" or so?
20:48:07 <ehird> Oh, the ACD must be 30"
20:48:15 <ehird> Anyway, it's larger than this display.
20:48:32 <ehird> bsmntbombdood: But really, if you look at the text on http://www.harmony-central.com/Test/wilson/two.jpg, a lot of those pixels will go to waste foro text.
20:48:37 <ehird> Still, the fractals will look nice. :P
20:48:52 <AnMaster> ehird, so will text with proper DPI settings
20:49:02 <ehird> You can completely disable hinting.
20:49:06 <AnMaster> ehird, also that high res concorde looks awesome
20:49:20 <ehird> bsmntbombdood: No ppi is too much!
20:49:25 <ehird> AnMaster: No, it's not quite that dense yet :P
20:49:27 <ehird> "This is a revised model of the original T220. Notable improvements include using only one power adapter instead of two"
20:49:43 <AnMaster> ehird, how much would be required for dropping AA?
20:49:44 <ehird> You need ~600 ppi to give up antialiasing.
20:49:50 <ehird> Same as in print, really.
20:50:27 * AnMaster wants those desktop bgs though
20:50:28 <ehird> An issue with the T221 is that doing awesome smooth 3D animations on it will be very hard on the graphics cards... (you need multiple)
20:50:34 <ehird> Which is such a shame.
20:50:47 <ehird> bsmntbombdood: Oh, and the refresh rate sucks:
20:50:52 <ehird> "The supported maximum refresh rates at native resolution depends on how many TMDS links are used. Single, double, and quad-link support 13, 25, 41 Hz refresh rates respectively. With reduced blanking periods single, double, and quad-TMDS-link can obtain 17.0, 33.72, and 41 Hz refresh. This model's internal refresh rate is always 41 Hz."
20:50:56 <AnMaster> ehird, but wouldn't that 30" ACD require that too?
20:51:02 <ehird> Only 41 Hz if you use four links.
20:51:06 <ehird> = moar graphics cards
20:51:22 <fizzie> Wikipedia has a messy "list of displays by pixel density" -- http://en.wikipedia.org/wiki/List_of_displays_by_pixel_density -- and admittedly the T220 with 204 ppi is the highest you can still call a "screen"; the rest are tiny device displays.
20:51:55 <ehird> Consumer-level cards can drive that today
20:51:59 <ehird> With two DVI links
20:52:09 <AnMaster> "The 9503-DG5 model had a native refresh rate of 48 Hz"
20:52:22 <AnMaster> "The IBM T221-DG5 was discontinued in June 2005."
20:52:23 <ehird> Yeah, but good luck finding one of them on the market
20:52:28 <ehird> All of them are discontinued
20:52:35 <ehird> You want it used anyway
20:52:40 <ehird> New, they're like $5,000
20:54:11 <AnMaster> "The Viewsonic VP2290b-3 is a rebadged version of this monitor.[citation needed]"
20:54:22 <ehird> 09:19:44 <ais523> suing IBM is widely regarded to be an incredibly stupid move
20:54:22 <ehird> No context required
20:54:25 <ehird> AnMaster: it is true
20:54:57 <ehird> Who knows, who cares. Too expensive. Buy used.
20:55:38 <AnMaster> ehird, anything like a 2D bitmap game would suck on it. But maybe scaling wouldn't look quite as bad as usual?
20:55:52 <ehird> Scaling would look fine.
20:55:57 <ehird> Each pixel is so small that it'd look perfect.
20:56:03 <ehird> No antialiasing required.
20:56:07 <fizzie> And you don't need more than one card if the card's good enough. (Admittedly all "good enough" cards I see in the local retailer's web page take up two slots and have four-digit prices, so...)
20:56:57 -!- augur has quit (Read error: 110 (Connection timed out)).
20:57:54 <ehird> 10:31:30 <fizzie> The sword alone can't stop!
20:58:47 <ehird> 10:36:36 <AnMaster> fungot, *prod*
20:58:47 <ehird> 10:36:37 <fungot> AnMaster: it's time you jumped off this mortal coil...
20:58:47 <fungot> ehird: the usual...test them. you can entertain us for awhile? yes, i'd have done something very brave! he's probably up north, to guardia!!! let's toast our land! now we'll have some peace! magus is a tad on the spooky side. our only hope.
20:58:47 <fungot> ehird: to the northwest of this cape. he took back the medal from the frog king. and i'd like to see that mystical sword for myself!
20:59:06 <ehird> 10:38:04 <fizzie> I think it's in the ocean palace, when you stick your stick into the mammon machine thing.
20:59:06 <ehird> This channel is PG-13.
21:00:10 <fizzie> No, no, you misunderstand, it's this glowy red thing.
21:03:01 <AnMaster> "<fungot> ehird: [...] our only hope." <-- why did this make me thing of star wars
21:03:01 <fungot> AnMaster: like, thanks princess. i'll take that under advisement!!!
21:03:20 <AnMaster> fungot, you are confused about gender
21:03:20 <fungot> AnMaster: cyrus! are you leaving! i'd forgotten how beautiful they are the evildoers? magus's lair! you brave! he's probably up north, to guardia!!! let's toast our land! now we'll have some peace! magus is a tad on the spooky side. our only hope.
21:03:51 <fizzie> Help me, Obi-Wan fungot, you are our only hope.
21:03:52 <fungot> fizzie: to the northwest of this cape. he took back the medal from the frog king. and i'd like to see that mystical sword for myself!
21:04:23 <fizzie> Apparently it was "my", not "our". But still.
21:04:58 <AnMaster> ehird, you never played chrono trigger?
21:05:20 <ehird> That PG-13 thing was a joke. But no.
21:05:28 <ehird> I do believe you didn't either until you asked fizzie what ct was. :P
21:07:48 <MizardX> I've played Chrono Trigger... and Chrono Cross
21:09:09 <ehird> I wonder if you can ditch /etc/hostname and just get the hostname from /etc/hosts.
21:10:33 <ehird> Generally the init system uses sethostname(), I guess.
21:10:48 <fizzie> The file itself is pretty superfluous, since the "gethostname" library function gets it from the kernel; as far as I know, the file's just used to sethostname on startup at some point.
21:14:00 <fizzie> Speaking of the host name, for some reason it annoys me when OS X automagically fiddles with the host name when I connect to interwebs via different wlans or gprsies. Makes the prompt all ugly. (Maybe I should just stick a fixed string in the prompt, though.)
21:15:43 * ehird tries to compile NetBSD's coreutil-type things on OS X
21:15:59 <ehird> fizzie: Doesn't do that for me.
21:16:00 <ehird> You can set the hostname, you know.
21:16:05 -!- Gracenotes has quit ("Leaving").
21:16:23 <ehird> [~]$ grep -r Bournemouth /etc
21:16:34 <ehird> /etc/hostconfig:HOSTNAME=Bournemouth
21:16:36 <ehird> Yes, it's "going away", whatever
21:16:53 <ehird> Then sudo hostname Bournemouth to kick it off, I guess
21:17:44 <fizzie> Fiddling with configuration files? How very not OS Xy. (I think I've set the "computer name" thing from the settings dialogs somewhere, though I can't be sure there's no hostname-setting somewhere in the networking setups.)
21:17:59 <ehird> I think the GUI edits that
21:18:32 <ehird> Anyway, then I'll just do sethostname(lookupdnswhateverthecurrentflavourofthemonthis("127.0.0.1"))
21:19:07 <ehird> Looks like I'll be writing my own init
21:19:30 <ehird> Can't be hard, can it
21:19:36 <ehird> Kernel's done all that pesky booting and all
21:22:30 <ehird> I wish I had an ext2 driver for OS X that can mount loopbacks.
21:23:47 <fizzie> I don't know anything about OS X's ext2 drivers, but if they are sensible enough, one would think that you could be able to just attach images with hdiutil.
21:24:06 <fizzie> "Raw disk images from other operating systems (e.g. .iso files) will be recognized as disk images and can be attached and mounted if OS X recognizes the filesystems." Well, I guess that depends.
21:24:31 <ehird> They don't actually exist, though.
21:24:44 <ehird> There isn't an ext2 driver that works in 10.5 as far as I can tell.
21:24:50 <ehird> Also, it's literally an .ext2
21:24:55 <ehird> Oh, great ... http://cvsweb.netbsd.org/bsdweb.cgi/src/bin/?only_with_tag=MAIN
21:25:04 <ehird> http://cvsweb.netbsd.org/bsdweb.cgi/src/usr.bin/?only_with_tag=MAIN
21:25:05 <AnMaster> <ehird> I wonder if you can ditch /etc/hostname and just get the hostname from /etc/hosts. <-- at least gentoo doesn't use /etc/hostname
21:25:41 <ehird> I guess weeding out the actual core utilities, getting them to build on Linux, and then working from that codebase is the thing to do
21:27:20 <ehird> Groan, I've forgotten how to use CVS.
21:27:58 <ehird> You have to do some kind of stupid login procedure that it somehow remembers the next command, don't you?
21:28:19 <ehird> cvs login anonymous blah blah blahh
21:28:20 <ehird> cvs checkout blah blah blah
21:28:39 <ehird> And it's really weird that it, you know, remembers your login the next command
21:28:42 <fizzie> Someone says ext2fsx's debug build works on Intel 10.5; nothing works on 10.6, though, and ext2fsx isn't the most trustworthy-looking piece of software there is. (sf.net "helpful review" #2: "Thanks for ruining my data. My drive is completely hosed.")
21:28:48 <AnMaster> ehird, I think it is stored in $HOME/.cvspass
21:28:56 <ehird> fizzie: Yeah, I saw that.
21:28:57 <AnMaster> rather than "remember to next command" only
21:29:51 <AnMaster> ehird, svn stores login stuff somewhere too. All VCS has to store stuff like push/pull/whatever urls and such
21:30:02 <AnMaster> well, all network enabled ones
21:30:32 <ehird> Yes, but the way it's "logging in" with CVS is weird.
21:30:53 <ehird> It's not "hey, remember my server details", it's modelled as "log in to the server, check out, and forget about it"
21:30:54 <fizzie> Oh, I've found the CVS separate-login-step strange too. It's a weird place to split the checkout/"clone" operation at.
21:31:06 <ehird> Groan, cvs directories everywhere
21:31:22 <ehird> [~/Junk]$ mkdir bin; cd bin; CVSROOT=anoncvs@anoncvs.NetBSD.org:/cvsroot cvs checkout -P src/bin
21:31:25 <ehird> Worked for me, FWIW
21:31:28 <ehird> Without a login step
21:31:56 <ehird> And it hasn't remembered anything; good
21:32:00 <ehird> #$NetBSD: Makefile,v 1.22 2007/12/31 15:31:24 ad Exp $
21:32:00 <ehird> #@(#)Makefile8.1 (Berkeley) 5/31/93
21:32:00 <ehird> SUBDIR=cat chio chmod cp csh date dd df domainname echo ed expr hostname \
21:32:01 <ehird> kill ksh ln ls mkdir mt mv pax ps pwd rcp rcmd rm rmdir sh \
21:32:01 <ehird> sleep stty sync test
21:32:01 <ehird> .include <bsd.subdir.mk>
21:32:08 <ehird> I'm not downloading your whole build infrastructure :P
21:32:20 <ehird> #$NetBSD: Makefile,v 1.12 2003/05/18 07:57:31 lukem Exp $
21:32:21 <ehird> #@(#)Makefile8.1 (Berkeley) 5/31/93
21:32:21 <ehird> .include <bsd.prog.mk>
21:32:24 <ehird> Nice Makefile, you cocks.
21:32:31 <AnMaster> /1 :pserver:anonymous@jsbsim.cvs.sourceforge.net:2401/cvsroot/jsbsim A
21:32:35 <ehird> At least OS X has bsdmake.
21:32:52 <AnMaster> (and I doubt anyone here is interested in jsbsim :P)
21:32:55 <ehird> "/Users/ehird/Junk/bin/src/bin/cat/../Makefile.inc", line 9: Malformed conditional ((${MKDYNAMICROOT} == "no"))
21:32:55 <ehird> "/usr/share/mk/bsd.init.mk", line 15: if-less endif
21:32:55 <ehird> bsdmake: fatal errors encountered -- cannot continue
21:32:57 <fizzie> That's what my ~/.cvspass looks like too; four repositories, all have just "A" (for "anonymous"?) as the secret.
21:33:04 <ehird> Oookay, my bsdmake has thingies.
21:33:09 <AnMaster> fizzie, there is some other one
21:33:14 <ehird> Then who was invalid error?
21:33:18 <AnMaster> /1 :pserver:cvsguest@cvs.flightgear.org:2401/var/cvs/FlightGear-0.9 AIbdZ,
21:33:31 <AnMaster> ehird, it probably needs the whole build tree?
21:33:39 <ehird> AnMaster: No, I have that /usr/share/mk file.
21:33:51 <ehird> I guess it needs to be NetBSD's for some reason.
21:34:05 <ehird> OS X shipping with a bsdmake file that its bsdmake trips over on
21:34:26 * ehird compiles cat manually
21:34:37 <ehird> Good code, incidentally.
21:34:44 <ehird> Apart from the mass of indentation.
21:34:57 <ehird> It reaches 7 levels of indentation.
21:35:04 <ehird> Function, for, if, if, if, and line continuation.
21:35:31 <ehird> Function, for, if, if, if, if, line continuation.
21:35:55 <ehird> [~/Junk/bin/src/bin/cat]$ ./cat -?
21:35:55 <ehird> ./cat: illegal option -- ?
21:35:55 <ehird> usage: cat [-beflnstuv] [-] [file ...]
21:36:05 <AnMaster> ehird, about cat. I met someone mad recently
21:36:05 <ehird> Oh dear, it has cat -v :-P
21:36:25 <AnMaster> ehird, no. Someone who claimed that a cat that didn't try to use mmap() when possible was basically shit
21:36:52 <AnMaster> (falling back on read() when mmap wasn't supported and/or the file didn't handle it, say, char device, standard input or such)
21:37:12 <ehird> Methinks csh will not be part of my core utilities.
21:37:16 <AnMaster> ehird, tell me, does that netbsd cat properly support -u?
21:37:25 <AnMaster> that is the only POSIX option for cat
21:37:28 <ehird> -u The -u option guarantees that the output is unbuffered.
21:37:34 <ehird> It may guarantee it by, say, doing nothing.
21:37:50 <ehird> Anyway, mmf. I'd prefer it had no options at all.
21:37:51 <AnMaster> ehird, freebsd one and gnu one doesn't do what you would epect
21:37:54 <ehird> Then again, compatibility is king...
21:38:04 <ehird> Most BSD utilities are almost identical.
21:38:30 <AnMaster> ehird, cat -e is like the only one I ever use. Not sure if *bsd cat -e is same as gnu cat -e
21:38:37 <ehird> [~/Junk/bin/src/bin/cp]$ cc cp.c utils.c -o cp
21:38:37 <ehird> Love how simple it is to compile
21:38:45 <ehird> AnMaster: -e and -v are evil.
21:39:08 <ehird> -e and -v are useful, but evil.
21:39:12 <ehird> -e is -v with printing endlines.
21:39:20 <ehird> "cat came back from Berkeley waving flags" — Rob Pike
21:39:31 <AnMaster> but yeah, those doesn't belong in cat
21:39:35 <ehird> cat should concatenate files; formatting them is outside its scope
21:39:37 <AnMaster> probably should be od or something like that
21:39:46 <ehird> It should be scrub
21:39:48 <ehird> or some other nice name
21:39:58 <AnMaster> ehird, scrub doesn't seem relevant here
21:40:01 <ehird> scrub for -v, scrub -n for newlines
21:40:03 <ehird> AnMaster: I made it up.
21:40:17 <ehird> It scrubs unprintables.
21:40:24 <ehird> scrub == cat -v; scrub -n == cat -e
21:40:29 <AnMaster> ehird, converts them to other chars instead
21:40:29 <ehird> Small utilities that do one thing. That's Unix.
21:40:46 <ehird> Come up with a better name.
21:40:51 <AnMaster> ehird, scurb should probably have a "just remove, not replace" option then
21:40:58 <ehird> But it shouldn't go in od, either, because that's a kitchen sink.
21:41:05 <AnMaster> ehird, oh and "unprintable". Clearly this needs to be locale and encoding aware
21:41:13 <ehird> AnMaster: DIE, FOUL DEMON!
21:41:15 <AnMaster> since in an UTF-8 locale it would differ
21:41:29 <ehird> Bringer of death and destruction!
21:41:48 <AnMaster> ehird, it should scrub the £ symbol btw
21:42:12 <ehird> Filthy non-mercans.
21:45:04 * ehird compiles NetBSD's csh for shits and giggles
21:45:11 <ehird> This is kinda pointless, I am, after all, on a BSD
21:45:22 <ehird> Still, it'll probably all run on Linux.
21:45:38 <ehird> csh is gross in more than usage; oh its code
21:46:19 <ehird> Yikes. csh doesn't build. Uh. Good.
21:49:31 <ehird> I wish X wasn't so mandatory.
21:50:21 <ehird> ("The X server has to be the biggest program I've ever seen that doesn't do anything for you." — Ken Thompson :P)
21:53:08 -!- FireFly has joined.
21:55:17 <AnMaster> ehird, hm I guess openoffice actually *does* something for you then
21:55:35 <ehird> Does rape count as doing something *for* you?
21:55:44 <ehird> It's a good question...
21:57:13 <ehird> It doesn't run on Linux, for one; it isn't open source, for one; it's unmaintained, for one; it doesn't support modern hardware, for one; it isn't backwards-compatible with X, for one.
21:57:44 <AnMaster> ehird, what is your opinion on gnustep?
21:58:14 <ehird> OS X's unpopular predecessor without the decent architecture? Maintained by GNU idiots?
21:59:37 -!- fax has joined.
22:00:46 <ehird> Wonder what distro and what hardware I should use to bootstrap.
22:02:53 <ehird> Oh, god, there is a site called "boycottboycottnovell.com"
22:02:55 -!- ehird has left (?).
22:02:57 -!- ehird has joined.
22:02:58 -!- ehird has left (?).
22:03:02 -!- ehird has joined.
22:05:49 -!- ehird has left (?).
22:05:54 -!- ehird has joined.
22:05:55 <fax> fungot: style
22:05:56 <fungot> fax: we are looking, but well behaved! crono!!! the monster who kidnapped the princess to the castle!
22:06:02 <fax> fungot: style
22:06:02 <fungot> fax: the knight spirit has the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the hero! hurrah to the
22:06:09 <fax> hurrah to the hero!
22:06:18 <ehird> Hurrah to the hero that sword alone can't stop.
22:09:05 -!- oerjan has joined.
22:09:18 <AnMaster> "boycottboycottnovell.com" <-- huh
22:10:22 <oerjan> boycottboycottboycottrecursion.com
22:11:01 <pikhq> endorseendorseendorseendorserecursion.com
22:11:34 <AnMaster> fizzie, that was a new one I think
22:11:36 <oerjan> read it hours ago (TM)
22:11:49 <AnMaster> oerjan, *tries to remember which theme it was*
22:12:20 <oerjan> but if steve summons enough of them...
22:13:11 <oerjan> they and cthulhu should form AAAS
22:13:30 <oerjan> Ancient Abominations Against Steve
22:13:54 <oerjan> actually make that AAAAAS
22:13:59 <ehird> Maybe I should finally give Gentoo a try...
22:14:03 <oerjan> Ancient Abominations And Alligators Against Steve
22:14:10 <ehird> But, ugh, what a waste of time.
22:15:23 <AnMaster> ehird, wait, you never tried it?
22:16:07 <ehird> Yeah, unless Gentoo somehow subverts logic and reason — or everyone who has said anything about Gentoo that I've read lied — I deducted that it is bad.
22:16:08 <AnMaster> ehird, also I'm pretty sure you wouldn't like it. Too many options
22:16:23 <ehird> I hope you are joking.
22:16:27 <AnMaster> ehird, as in, you have three cron implementations to choose from in the standard distro
22:16:37 <ehird> Regardless, I'd bother trying it if it wasn't for the whole "LOL 24 HOUR INSTALL PROCESS"
22:16:48 <ehird> "Look at the massive size of our source-based ecocks"
22:17:01 <ehird> AnMaster: The simplest and smallest!
22:17:14 <AnMaster> ehird, ah you want to turn off most useflags then
22:17:24 <fax> hurrah to the hero!
22:17:35 <AnMaster> needless to say I have more than usual ammount
22:17:45 <AnMaster> USE="3dnow acpi ccache sse sse2 pic nptl glep -gnome mmx pcre unicode objc bash-completion acl 3dnowext caps emacs cairo logrotate vorbis jpeg2k openexr fontconfig mozdevelop nsplugin ieee1394 lm_sensors fbcon dvdr sndfile javascript -java mysqli iconv gmp bzip2 exif cdb gd curl ogg truetype gdbm expat flac mad audiofile mng lcms idn dri mono scanner sqlite ppds -eds tcl tk nptlonly -ldap usb foomatic
22:17:45 <AnMaster> db tiff -esd -oss xpm -hal imlib -xml utempter idea mbox pdf mmxext physfs qt3support ipv6 xcomposite -kerberos kqemu zsh-completion gnutls iproute2 joystick dbus rle -apache2 fastcgi kdehiddenvisibility kdeenablefinal nodrm lyx loop-aes -arts geoip -branding -libnotify mmap -mysql webdav-serf sasl -bluetooth -openmp cddb pg-intdatetime -accessibility -ssse3 -consolekit -php"
22:18:05 <oerjan> AnMaster: did you have to do that just as i was about to chastise fax for spamming
22:19:00 <ehird> it was two lines nobody really cares about
22:19:00 <ehird> also, -libnotify? why?
22:19:01 <AnMaster> ehird, that is just the global use flags
22:19:12 <AnMaster> ehird, there are package specific ones in another file
22:19:21 <AnMaster> like category/package foo -bar
22:19:41 <fax> oerjan what the hell are you talking about
22:19:54 * pikhq wonders if Gentoo still supports using uclibc...
22:19:56 <ehird> hmm, i should write a libnotify client/server/whatever thing that uses dwm's status
22:20:06 <AnMaster> ehird, I didn't see the point of libnotify
22:20:17 <oerjan> fax: <fax> hurrah to the hero!
22:20:34 <pikhq> Oh, god. A 2005.1 profile?
22:20:37 <ehird> Lets you know when you have been emailed / IMed / name mentioned on IRC.
22:20:38 <pikhq> I'm going to say "no".
22:20:46 <ehird> tbh I'll probably bootstrap from arc
22:20:54 <AnMaster> pikhq, I remember back before eselect
22:21:23 * pikhq finds eselect tolerable, but would something better.
22:22:11 <AnMaster> a bit slow, always wonder how they managed that
22:22:18 <pikhq> It's a shell script.
22:22:59 <AnMaster> pikhq, opinion on pkgcore? And on paludis (spelling?)
22:23:39 <pikhq> Never tried either.
22:24:11 <AnMaster> pikhq, paludis depends on boost you know? And is written (from what I have heard) in rather bad style. As in "worst of C++"
22:24:17 <fax> fungot: style
22:24:17 <fungot> fax: you! take! we find! that sword alone can't stop, crono! are you leaving!
22:24:33 <fungot> AnMaster: by thy leave, crono?!! you brought back my cat! thank you, crono! are you leaving!
22:24:34 <fax> what's the most esoteric language
22:24:36 <ehird> You know that boost makes using C++ not make you want to kill yourself too much, right?
22:24:40 <ehird> At least if you're good at self-delusion.
22:24:45 <ehird> Sure, internally... but...
22:25:13 <AnMaster> ehird, well, sure, but a package manager depending on something so prone to ABI breakage
22:25:22 <AnMaster> and, well, it is unclean in other parts
22:25:35 <ehird> ABI breakage! Hooray, static binaries.
22:25:42 <pikhq> Not that Portage is much better, AnMaster. It uses Python...
22:25:46 <AnMaster> oh and the main dev is quite... rude against everyone who disagrees with him
22:25:58 <ehird> Static binaries — never worry about your package manager being broken!
22:25:59 <AnMaster> pikhq, true, but generally that breaks less
22:26:15 <AnMaster> ehird, arch used to have pacman.static but iirc it was dropped
22:28:04 <ehird> Heck, you can even manage your packages from a linux boot floppy. :P
22:28:13 <ehird> By which I mean a really pathological one.
22:28:17 <ehird> Otherwise that's nothing special.
22:29:00 <ehird> IDEA: Statically link the kernel into every binary!
22:29:14 <ais523> that would only work if you could easily hotswap kernels
22:29:17 <fizzie> Hurrah, hurrah. (It's a combination of "Hurrah to the Hero and Guardia!" and "Peace at last, thanks to the Hero! Hurrah for the Hero! Hurrah for Guardia!".)
22:29:24 <ais523> heh, every binary in its own VM would be decent for security
22:29:57 <ehird> ais523: Total portability, though!
22:30:09 <fizzie> The Perl script again doesn't get stuck in a loop: http://pastebin.com/m73626919
22:30:10 <pikhq> And would actually work tolerably on a non-x86 system.
22:30:13 <AnMaster> ais523, a bit of pain for usability
22:30:15 <ais523> no, because you'd need some OS-independent way to run the kernels
22:30:26 <ais523> clearly, you need to compile the bootloader in too
22:30:44 <ehird> ais523: And the CPU microcode.
22:30:49 <ehird> And the CPU microcode interpreter.
22:30:56 <AnMaster> ais523, you would need an universal binary to support different platforms
22:30:58 <ais523> AnMaster: multiboot as in simultaneous multiple boot
22:31:02 <ais523> ehird: and the VHDL for the hardware
22:31:03 -!- Oranjer has joined.
22:31:08 <ehird> AnMaster: You just include the hardware.
22:31:14 <ehird> ais523: and the VHDL emulator
22:31:20 <ehird> which is a linux binary
22:31:42 <ais523> you get your computer to build new computers to run the applications on
22:31:56 <ais523> incidentally, this conversation is a good reflection on Feather
22:32:29 <AnMaster> ais523, how is it a good reflection on Feather?
22:32:45 <ais523> AnMaster: the infinite regress of emulation layers
22:32:48 <ais523> Feather manages to simulate that
22:33:00 <ais523> by retroactively adding more stages to the regress whenever they'd ever become relevant
22:33:05 <fizzie> "hurrah to the hero, he might be the one to bring forth an immense evil... ...humans make them that way."
22:33:48 <ais523> AnMaster: no, busy preparing for devnull
22:33:52 <ais523> are you planning to play?
22:34:04 <fax> hurrah to the hero!
22:34:30 <ehird> devnull? Rings a bell.
22:35:17 <AnMaster> ais523, plus I never liked playing against people
22:35:22 <ais523> ehird: NetHack tournament
22:35:26 <AnMaster> ais523, as in, I much prefer single player
22:35:38 <AnMaster> and that much lag is implausible
22:35:42 <ais523> it thinks it's the longest-running Internet gaming tournament in existence, and nobody has sent it a counterexample yet
22:35:55 <ehird> Also, it thinks nothing.
22:35:59 <AnMaster> ais523, counterexample of what?
22:36:23 <ais523> <devnull> /dev/null has been holding an annual NetHack Tournament, beginning at midnight on Halloween, since 1999. This appears to make /dev/null/nethack the longest running gaming tournament on the Internet; folks who've looked into this have told us that they've found two other tournaments claiming this title, one at 6 years old and the other at 4 years old (as of the summer of 2008), so this may very well be the case.
22:36:42 <AnMaster> ais523, oh should that be "sent in" rather than "sent it"?
22:36:55 <ehird> ais is anthropomorphising it.
22:36:58 <ais523> AnMaster: I was anthropomorphising inappropriately
22:37:01 <ehird> ais523: well, it certainly isn't the oldest, and it certainly isn't the one with the most games played
22:37:10 <ehird> AnMaster: don't worry, it's very awkward and confusing to me too
22:37:32 <ais523> ehird: what older one do you know of?
22:37:38 <ais523> I'm sure they'd be happy to find a counterexample
22:37:53 <ehird> Of course people have organised tournaments before that.
22:37:58 <ehird> They might not still exist, though.
22:38:16 <AnMaster> ehird, the point is not when, but for how long
22:38:21 <ehird> e.g. Quake II tournaments have almost certainly been organised before 1999; I don't know if any still exist
22:38:33 <ehird> [21:37] ehird: ais523: well, it certainly isn't the oldest, and it certainly isn't the one with the most games played
22:38:44 <ehird> I was saying that it only has a chance of winning in one metric.
22:38:57 <AnMaster> but that was the relevant metric for this discussion
22:39:12 <ehird> I was just noting.
22:39:51 * AnMaster realises he would need a new computer to handle it
22:40:12 <AnMaster> since I doubt I could fit in high end enough graphics card
22:41:14 <ehird> Well, AGP was the thing when the T221 was around...
22:41:52 <ehird> Anyway... get two Radeon 5870s... shit iwll scream.
22:42:01 <ehird> (And a new computer to fit them :P)
22:42:24 <AnMaster> ehird, yeah I bust my budget at the screen alone already
22:43:08 <AnMaster> oh well, one can always dream I guess...
22:43:13 <ehird> 5870 costs about $400
22:43:33 <ehird> So I'd say $3,300 total expense could get you two of them and a T221.
22:47:02 <ehird> No T221s seem to be for sale on the int'webs.
22:48:46 <ehird> Very high-resolution display, IBM, circa 2001.
22:49:17 <ehird> If you're lucky, you can get 41 Hz refresh out of it.
22:49:33 <ehird> Less DVI links? 13 Hz.
22:49:45 <ehird> (13, 25, 41 for single, double and quad DVI links)
22:50:15 <ehird> Well, yeah, but that one's hard to get.
22:50:23 <ehird> Like, I don't know if anyone had it.
22:50:30 <ehird> Reading it again, it would be easy.
22:50:33 <ehird> Since it's just a regular update.
22:50:38 <ehird> So 48 Hz, pretty good. But still.
22:50:49 <ehird> That's 48 fps max.
22:50:52 <AnMaster> ehird, my normal TFT runs at 60 Hz iirc
22:51:11 <ehird> Well, apart from this oddball.
22:51:16 <ehird> "On 19 March 2002, IBM announced lowering the price of IBM T221 from US$17,999 to US$8,399."
22:51:31 <ehird> So the guy who got it for $2,500 in 2004 is one lucky fucker.
22:53:08 <ehird> http://blogs.codehaus.org/people/topping/archives/000856_life_with_the_ibm_t221.html
22:53:35 <ehird> http://codehaus.org/~topping/screen.JPG
22:53:37 <ehird> is a screenshot of his work space
22:53:42 <ehird> Notable is that TINY FONT in Eclipse.
22:53:51 <ehird> Wait, not Eclipse.
22:54:32 <ehird> Oh, he got it for $3,000
22:55:05 <ehird> "I typically run by day it at max resolution, and by night, reconfigure it to 1920x1200. This is basically uses a 2x2 square of native screen for every image pixel and is a more standard resolution at about 100dpi (good for tired eyes!)"
22:55:10 <ehird> You could use that to play games.
22:55:22 <ehird> Good res for such a smooth scaling.
22:55:37 <ehird> But yeah: Ow, tiny!
22:55:40 -!- augur has joined.
23:05:40 <oerjan> and augur kills the channel again
23:06:40 <ehird> Hardware should be illegal!
23:07:41 <ehird> ps ais523 http://www.emacswiki.org/pics/static/TabsSpacesBoth.png
23:07:56 <ais523> ehird: and that's on emacswiki?
23:07:56 <oerjan> if we're in the matrix, and the matrix is recursive, then maybe there is no hardware
23:08:02 <ais523> tell emacs not to use tabs and spaces by default
23:08:24 <ehird> ais523: that doesn't fix other people's code.
23:08:46 <fizzie> It's on the page where they tell you how to configure it to use tabs for indentation, spaces for alignment. (At least; it might be on other pages too.)
23:08:59 <oerjan> it's virtuals all the way up!
23:09:03 <Oranjer> oerjan, how could the matrix be recursive?
23:09:05 <ehird> Which isn't the same as indenting with both.
23:09:22 <fizzie> (http://www.emacswiki.org/emacs/SmartTabs, that is.)
23:09:23 <oerjan> Oranjer: by being simulated on another matrix, duh
23:09:41 <Oranjer> that's merely nested, perhaps infinitely
23:09:59 <ehird> fizzie: ais523's reply to that is "tabs are 8 spaces wide always because i said so also everybody thinks this way so there"
23:10:08 <oerjan> maybe it's corecursive instea
23:10:21 <Oranjer> I thought you meant "recursive" as "matrix A contains matrix B, matrix B contains matrix A"
23:11:02 <oerjan> um no, and recursion doesn't have to be identical levels
23:11:19 <oerjan> fac n = n * fac (n-1) never repeats an n between levels
23:11:21 <Oranjer> well, there could be any number of matrixes between them
23:11:53 <oerjan> (and i know i'm leaving out the base case)
23:12:00 -!- FireFly has quit ("Later").
23:12:21 -!- FireFly has joined.
23:12:29 <ehird> "Rotate your widescreen 90 degrees." fuck people who say this, i can't do that :<
23:12:45 <Oranjer> well then, oerjan, what do you mean by recursive matrices?
23:13:24 <oerjan> Oranjer: i think corecursive fits my intuition better, actually. each matrix is simulated inside another, infinitely.
23:13:44 <AnMaster> <fizzie> It's on the page where they tell you how to configure it to use tabs for indentation, spaces for alignment. (At least; it might be on other pages too.) <-- That is The Right Way
23:13:53 <Oranjer> there's no "reality" that the first matrix is in?
23:14:02 <ehird> Oranjer: you are having real troubles with his joke
23:14:08 <ehird> anyway sure there is PER MODAL REALISM
23:14:11 <oerjan> there is no first matrix
23:14:25 <ehird> AnMaster: Yes, but it's a bitch because of idiots in any language but c.
23:14:43 <AnMaster> "because of idiots in any language but c"?
23:15:02 <oerjan> unless you count from the lowest level, where we are (unless our universe contains a matrix for another somewhere, which might be possible)
23:15:29 <oerjan> hm then it could be infinite both ways
23:15:31 <ehird> AnMaster: protip: if the nesting seems strange, PARSE IT AGAIN
23:16:03 <Oranjer> I presumed that's what you meant by infinite
23:16:11 <Oranjer> also, it reminds me of this short story: http://qntm.org.nyud.net:8090/?responsibility
23:16:19 <AnMaster> ehird, i seemed sane. Just unehirdic ;P
23:16:21 <ehird> nyud.net? seriously/
23:16:24 <AnMaster> anyway no I can't figure it out
23:16:43 <ehird> AnMaster: no, it's more fun having you trying to figure it out
23:16:52 <ehird> URL without nyud: http://qntm.org/?responsibility
23:17:10 <ehird> Nope, I've lost nothing
23:17:29 <AnMaster> ehird, yes you did. Someone stole that while you were looking the other way!
23:17:47 <ehird> 1. Reincarnation, fucker
23:17:47 <ehird> 2. NO DESIRES, FUCKER
23:18:04 <ehird> Buddhism is awesome because you can cheat at it!
23:18:12 <ehird> Just never become enlightened, ever, and you'll be reincarnated eternally.
23:18:35 -!- Pthing has joined.
23:19:17 <oerjan> AnMaster: hey i was remembering that story too
23:19:22 <AnMaster> ehird, I think you can be even when enlightened. Look at Dalai Lama
23:19:26 <ehird> AnMaster didnt point to it
23:19:36 <ehird> AnMaster: What about him?
23:19:41 <oklopol> ehird: how's that cheating?
23:19:45 <Oranjer> http://qntm.org/?responsibility
23:19:49 <ehird> In religious terms, the Dalai Lama is believed by his devotees to be the rebirth of a long line of tulkus, who have chosen to be reborn in order to enlighten others
23:19:53 <ehird> the lama isn't enlightened, therefore
23:20:00 <ehird> presumably, just close
23:20:17 <ehird> oklopol: because the cycle is painted as suffering; you're meant to become enlightened, where you die the typical atheist no-more-you death
23:20:28 <AnMaster> ehird, pretty sure enlightenment is like god mode (light edition) or such :P
23:20:28 <ehird> well, rather, if you're enlightened when you die
23:20:51 <ehird> If you're enlightened, and you die, that's it, dead.
23:21:51 <ehird> http://shii.org/afterlife which, while old and inaccurate for other religions, was written by a Buddhist.
23:22:04 <Oranjer> http://en.wikipedia.org/wiki/Sa%E1%B9%83s%C4%81ra
23:22:18 <Oranjer> it's the Buddhist concept of the cycle of life, death, etc.
23:22:21 <ehird> So, cheating at Buddhism = FUCK YEAH! Reincarnation!
23:22:38 <ehird> I'm being silly, Oranjer.
23:22:58 * oerjan notes that MaxChaplin in the comments thought of the same solution as he did
23:24:17 -!- Sgeo has joined.
23:24:35 <AnMaster> oerjan, <AnMaster> oerjan, what story?
23:24:47 <ehird> He meant to target Oranjer.
23:24:55 <ehird> Or perhaps I, who decrufted his link.
23:25:06 <Oranjer> http://qntm.org/?responsibility
23:27:04 <ehird> Sam Hughes is amazing.
23:28:26 <ehird> 2009-07-31 12:19:56 by Sam:
23:28:26 <ehird> I don't know what you people are talking about. There is a top reality. Because I say so. It's my story.
23:28:28 <ehird> Oranjer: sam hughes = qntm.org
23:28:41 <Oranjer> okay, I just found that out with google
23:34:57 <ehird> http://www.podtycoon.com/shutdown/
23:35:42 <ehird> I almost wished it were true.
23:36:20 <Oranjer> I actually had the theory that we were living in a documentary about the real Bush's life
23:36:32 <ehird> Go back to the page.
23:36:36 <ehird> Or you'll be randomly assigned.
23:36:52 <ehird> Shut up, you fool! Go back!
23:40:03 <ehird> "A Formalization of Darcs Patch Theory Using Inverse Semigroups"
23:42:22 <AnMaster> ehird, done it. Selected online
23:42:41 <ehird> There'd be an opportunity to argue for AI rights later.
23:42:48 <ehird> And it leaves the possibility of virtual simulations.
23:42:59 <ehird> Plus it's the only option which lets you contact the real world.
23:45:35 <fizzie> It is the most popularest thing. Though I have to wonder: why Finland?
23:45:48 <Oranjer> hey, I also selected Online!
23:45:49 <ehird> Maybe the author lives there.
23:46:02 <fizzie> Not according to his life-story page; http://www.fordfam.com/matthew/matthew_life.html
23:46:21 <ehird> Anyway, because it's a techno-ish country?
23:46:23 <Sgeo> I must not have been an AC. It just said "Poof!"
23:46:37 <ehird> Online is basically the best idea.
23:46:40 <Sgeo> I guess that's what it does for bots, to keep them thinking it's a joke
23:46:41 <ehird> It will redirect, and you're—
23:46:50 <fizzie> fungot: What did you pick?
23:46:50 <fungot> fizzie: frog will do. that frog's hand! you got the broken! the mountain of woe. it's likely that dalton came from the laboratories to the west?... yes! well then rest and relax! huh?! well, remember that you can log in anywhere on the world map! need a brief weapons and items seminar?
23:46:55 <ehird> Sgeo must be a bot now.
23:46:58 <fizzie> I don't think there was a "frog" option.
23:47:09 <ehird> It must have stayed at Poof!, instead of waiting a second or two and it happening.
23:47:37 <ehird> Online is the best option, although if you just want infinite bliss, dream or an appropriate religion's afterlife is good.
23:47:43 <ehird> Oranjer: Don't be silly!
23:47:53 <ehird> Of course we're not. You can hear me, right?
23:47:59 <ehird> The page is clearly just a joke. Pretty stupid one too.
23:48:03 <ehird> Why are we wasting our time on this?
23:48:16 <ehird> Come on, it's retarded! Like a futurist's wet dream! HURFDURF
23:48:21 -!- coppro has joined.
23:48:22 <ehird> Yeah, I was linking to it because it's REEEEEEEETAAAAAAARDEEEEEEEEEEED
23:48:23 <Sgeo> I think ehird's bot is acting the way it thinks ehird would pretend to be a bot
23:48:25 <Oranjer> because you're a bot tasked with finding the rest of the AC's!!!
23:48:59 <ehird> "The bots are programmed to mock it" is so stupid because it DEFLECTS REAL CRITICISM; like YOUR MOM WOULD.
23:48:59 <ehird> Huuurrrrrrrrrrrrrrr
23:49:15 <ehird> And, furthermore, "And, furthermore, "And, furthermore, "Stack overflow
23:50:16 <Oranjer> "the devil put the fossils in the ground, and constructed the pre-christian religions"
23:50:51 <Oranjer> I don't believe bots aren't real people
23:51:07 <ehird> Why do you think I don't believe bots aren't real people?
23:51:15 <ehird> What makes you feel like ha?
23:51:35 <Oranjer> I never said anything about your opinion on bots
23:51:49 <ehird> Can you elaborate on I never said anything about your opinion on bots?
23:51:52 <Oranjer> the "nooo ehird" was "oh no ehird you're a bot"
23:52:07 <ehird> I'll do the talking here.
23:52:22 <Oranjer> you asked me why I thought you don't believe bots aren't real people, and I responded by saying that I never said such a thing
23:52:39 <ehird> Why do you think you feel you asked me why I thought you don't believe bots aren't real people, and I responded by saying that I never said such a thing?
23:52:58 <Sgeo> ehird's been replaced with ELIZA
23:53:10 <ehird> Why do you think that ehird's been replaced with ELIZA? Something from your childhood?
23:53:25 <AnMaster> <ehird> I'll do the talking here. <-- fail. should be "asking"
23:53:39 <ehird> AnMaster: I'll do the talking here.
23:53:48 <ehird> Why do you feel AHHHHHHHHHH?
23:53:56 <ehird> Please, tell me more about save yourself!.
23:54:10 <ehird> Why do you say hahahaha?
23:54:22 <Oranjer> fungot, did you and ehird switch consciousnesses?
23:54:23 <fungot> Oranjer: to the northwest of this cape. he took back the medal from the frog king. and i'd like to see that mystical sword for myself! geez!
23:54:26 <ehird> AnMaster: Permission denied
23:54:32 <ehird> You are not in /etc/sudoers
23:54:35 <ehird> This incident will be reported.
23:54:47 <AnMaster> ehird, no one reads system logs anyway (except me)
23:55:03 <ehird> It sends an email to root, actually.
23:55:19 <ehird> Also, good point. Maybe I should disable logs. :P
23:55:30 <AnMaster> ehird, no. Good for when things break
23:55:42 <ehird> Some logs could do with disabling though...
23:56:01 <AnMaster> ehird, but most people don't routinely check for non-whitelisted log entries
23:56:46 <ehird> Grr, I wish there was a good minimal coreutils replacement.
23:56:50 <AnMaster> ehird, list of regex, anything not matching in logs is reported to me
23:56:59 <ehird> Instead of digging through BSD source trees...
23:57:41 <ehird> Busybox is not minimalist, it's useless. :P