00:00:39 -!- sebbu2 has quit (Read error: Connection reset by peer). 00:01:08 -!- sebbu has joined. 00:02:00 the number of \'s between each " goes as 2^n-1 00:02:00 !haskell :t Data.Function.fix (concatMap show) 00:02:00 now what 00:02:00 !echo hi 00:02:00 hi 00:02:02 Data.Function.fix (concatMap show) :: [Char] 00:02:28 that one _should_ have done something similar, characterwise 00:02:42 -!- comex has quit (Excess Flood). 00:02:59 but it doesn't... because show on a Char isn't lazy enough to give the initial ' without looking at the character first... 00:03:16 but if we cheat a bit we can see what that should have looked like 00:03:35 -!- sebbu has quit (Changing host). 00:03:35 -!- sebbu has joined. 00:03:35 !sh while true; do echo -n y; done 00:03:40 (That was dumb) 00:03:58 !haskell take 100 $ Data.Function.fix (('\'':).tail.concatMap show) 00:04:01 "'\\'''\\\\''\\'''\\'''\\'''\\\\''\\\\''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\\\''\\'''\\''" 00:04:08 Not sure *shrugs* 00:04:13 Probably something to do with how it outputs that. 00:04:15 !haskell putStrLn . take 100 $ Data.Function.fix (('\'':).tail.concatMap show) 00:04:16 '\'''\\''\'''\'''\'''\\''\\''\'''\'''\\''\'''\'''\'''\\''\'''\'''\'''\\''\'''\'''\'''\\''\\''\'''\'' 00:04:20 -!- hagb4rd has quit (Ping timeout: 255 seconds). 00:04:25 It might happen to work if it flushes. 00:04:30 -!- hagb4rd has joined. 00:04:46 !haskell Data.Function.fix (('\'':).tail.concatMap show) 00:04:59 "'\\'''\\\\''\\'''\\'''\\'''\\\\''\\\\''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\\\''\\'''\\'''\\\\''\\\\''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\\\''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\\\''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\\\''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\'''\\'''\\'''\\\\''\\\\ 00:05:03 Gregor: weird that it would flush better with that than with an infinite list. maybe it also has something to do with speed of printing... 00:05:27 maybe it normally times out before getting far enought that ghc needs to flush 00:05:32 *-t 00:05:50 where is lambdabot? 00:06:03 !haskell Data.Function.fix ('a':) 00:06:21 -!- comex has joined. 00:06:27 that didn't work :( 00:06:34 !echo hi 00:06:45 well 00:06:47 !help 00:06:47 help: General commands: !help, !info, !bf_txtgen. See also !help languages, !help userinterps. You can get help on some commands by typing !help . 00:06:53 there is no echo? 00:07:19 !help userinterps 00:07:19 userinterps: Users can add interpreters written in any of the languages in !help languages. See !help addinterp, delinterp, show | !userinterps. List interpreters added with !addinterp. 00:07:26 !userinterps 00:07:27 Installed user interpreters: acro aol austro bc bct bfbignum brit brooklyn bypass_ignore bytes chaos chiqrsx9p choo ctcp dc decide drawl drome dubya echo eehird ehird elmer fudd google graph gregor hello id jethro kraut lperl lsh map num ook pansy pi pikhq pirate plot postmodern postmodern_aoler prefixes python redneck reverse rimshot rot13 rot47 sadbf sanetemp sfedeesh sffedeesh simplename slashes svedeesh swedish ustemp valspeak warez wc yodawg 00:07:30 night → 00:07:32 Vorpal: ^ echo 00:08:23 Vorpal: netsplat a moment ago, after not responding for a while 00:11:47 -!- Vorpal has quit (Ping timeout: 246 seconds). 00:11:48 -!- hagb4rd has quit (Ping timeout: 255 seconds). 00:12:19 -!- hagb4rd has joined. 00:15:17 !haskell Data.Function.fix (\x -> "LOOK BEHIND YOU " + x) 00:15:44 !haskell Data.Function.fix (\x -> "LOOK BEHIND YOU " ++ x) 00:16:38 !haskell Data.Function.fix show 00:16:53 "\"\\\"\\\\\\\"\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ 00:18:30 maybe it on the contrary works because fix show is _slow_... 00:18:56 hm... 00:19:11 !haskell Data.Function.fix ("\\\""++) 00:19:40 or maybe EgoBot just likes Lymia. 00:19:43 !echo hi 00:20:10 * hagb4rd chekcs his codepage configuration 00:20:17 !haskell Data.Function.fix (\x -> show $ "LOOK BEHIND YOU " ++ x) 00:20:31 "\"LOOK BEHIND YOU \\\"LOOK BEHIND YOU \\\\\\\"LOOK BEHIND YOU \\\\\\\\\\\\\\\"LOOK BEHIND YOU \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"LOOK BEHIND YOU \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"LOOK BEHIND YOU \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\"LOOK BEHIND YOU \\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\\ 00:20:35 aha 00:21:04 i'm wondering if there may be some kind of balance... it needs to print enough, and _then_ time out? 00:22:27 would be kool to see just the parts with high information density 00:22:27 !haskell "testing... " ++ Data.Function.fix id 00:22:30 *** Exception: stack overflow 00:22:35 hmph 00:23:17 Gregor: it seems EgoBot then went on to try running it as a _module_ afterwards... 00:23:54 oh well 00:27:05 -!- augur has quit (Remote host closed the connection). 00:37:08 -!- CakeProphet has quit (Ping timeout: 260 seconds). 00:38:59 -!- CakeProphet has joined. 00:39:00 -!- CakeProphet has quit (Changing host). 00:39:00 -!- CakeProphet has joined. 00:48:06 -!- yorick has quit (Ping timeout: 240 seconds). 00:48:24 -!- yorick has joined. 00:48:49 -!- augur has joined. 00:49:21 -!- copumpkin has quit (Read error: Connection reset by peer). 01:03:09 * pikhq_ looks for good compile-time undefined behavior 01:03:26 -!- FireFly has quit (Quit: swatted to death). 01:21:11 Fuckit. */ 01:22:12 -!- lambdabot has joined. 01:30:11 -!- oerjan has quit (Quit: leaving). 01:38:34 -!- elliott has joined. 01:42:07 x 02:03:48 -!- copumpkin has joined. 02:08:14 -!- pizearke has quit (Quit: quuiiiitttt!). 02:08:33 -!- pizearke has joined. 02:54:37 elliott: you're male most of the time, right? 02:54:57 -!- Adamfyre has joined. 02:55:40 tswett: That's... a technically true statement :P 02:56:55 So, my question is true. 02:57:25 Does that mean that it's a good, precise question, whose answer is likely to enlighten rather than confuse? 02:57:38 This sure is a conversation. 03:00:47 -!- Adamfyre has quit (Quit: AndroidIrc Disconnecting). 03:06:16 -!- sebbu has quit (Read error: Connection reset by peer). 03:06:44 -!- sebbu has joined. 03:11:57 -!- zzo38 has joined. 03:30:38 -!- pikhq has joined. 03:33:46 It's just OUCH? 03:34:07 -!- pikhq_ has quit (Ping timeout: 260 seconds). 03:34:35 tswett: You're just OUCH most of the time, right? 03:40:36 -!- azaq231 has joined. 03:41:04 -!- azaq23 has quit (Ping timeout: 246 seconds). 03:46:31 -!- piz2 has joined. 03:50:13 -!- pizearke has quit (Ping timeout: 240 seconds). 03:51:11 -!- CakeProphet has quit (Ping timeout: 264 seconds). 03:54:27 -!- oerjan has joined. 03:56:45 !unlambda `c.i 03:56:45 i 03:57:09 !unlambda `c.i.t 03:57:10 i 03:57:18 !unlambda ``c.i.t 03:57:18 it 03:57:36 !unlambda ```c.i.t.y 03:57:36 itt 03:58:31 !unlambda `````.S``c.u.c.e.s.s 03:58:32 ./interps/unlambda/unlambda.bin: file /tmp/input.32133: parse error 03:58:58 !unlambda ````.S``c.u.c.e.s.s 03:58:58 ucSces 03:59:46 !unlambda ``````c`.S.u.c.e.s.si 03:59:46 Success 04:00:07 zzo38: was that your solution too? 04:01:29 !unlambda `````c`.S.u.c``c.e.si 04:01:30 ./interps/unlambda/unlambda.bin: file /tmp/input.32431: parse error 04:01:36 argh 04:01:48 !unlambda ````c`.S.u.c``c.e.si 04:01:49 Sucescs 04:01:56 oops 04:02:14 sucesces 04:02:19 so close :P 04:02:25 What are you trying to do, BTW? 04:02:51 zzo38 mentioned he'd got this golf down to 21 in unlambda 04:03:05 ah 04:03:07 oerjan: That was not quite my solution but it was similar to that 04:04:03 hm 04:04:22 !unlambda ``````c`c`.S.u.c.e.si 04:04:22 ./interps/unlambda/unlambda.bin: file /tmp/input.32699: parse error 04:04:26 !unlambda `````c`c`.S.u.c.e.si 04:04:27 Succes 04:04:31 darn 04:06:39 zzo38: oh wait there should be a . at the end? 04:06:47 oh, no 04:08:27 !unlambda ```c```c`.S.u.c.e.si 04:08:27 Success 04:08:30 yay 04:08:56 Post to anarchy golf if you did it good. 04:09:33 i will. what's the open code-statistics? 04:10:24 I always like to select that option. It means it tells you how many binary/alphanumeric/symbols 04:10:38 ok i did too 04:11:01 (Look below, under the "Statistics" column for the summary of solutions, that is what those numbers means.) 04:11:09 yay my first golf submission (afair) 04:14:51 huh someone manage brainfuck in 18 bytes 04:14:53 *d 04:15:08 what the hell are you doing, oerjan 04:15:18 oerjan: It is a BFI-specific code probably. 04:15:18 augur: golfing 04:15:24 golfing? 04:15:37 It takes advantage of unintentional features in BFI. 04:15:37 augur doesn't know about code golf? 04:15:41 no 04:15:42 i dont 04:15:43 aha 04:16:17 lol 04:16:24 i mean, i might, but 04:16:35 oerjan: btw if you submit with a web browser it appends \r\n I think... or wait, does it strip all whitespace 04:16:36 I forget 04:16:38 zzo38: hm it cannot be just ! with input, that would have been shorter 04:16:52 I do know that it uses \r\n newlines with the web form though so if you have any newlines file submission is best 04:16:58 (As you can see, it is obviously not a proper brainfuck, since it has 1 binary and 7 letters (the word "Success" is in there) 04:17:02 elliott: erm i used the form, no newlines 04:17:06 zzo38: hm it cannot be just ! with input, that would have been shorter 04:17:10 bfi allows self-modifying code or something 04:17:18 aha 04:17:18 elliott: Yes it does do that. It does not strip any whitespaces 04:17:28 zzo38: but does it append \r\n with the web form? 04:17:31 ISTR it does 04:17:35 so the web form always costs you two bytes 04:17:37 not sure though 04:17:42 maybe not 04:17:48 elliott: Yes it does append \r\n when using the HTML form 04:17:55 elliott: Yes it allows accessing the code-space (and modifying it) 04:18:02 oerjan: you should make a file locally and upload it to save two bytes 04:18:15 But it does not append \r\n to the end. 04:18:22 It only converts newlines to \r\n 04:18:28 oh 04:18:31 but you said 04:18:32 elliott: Yes it does append \r\n when using the HTML form 04:18:32 ? 04:18:46 I know, I made a mistake in the typing sentence 04:18:49 I didn't mean that. 04:18:57 OK 04:18:58 elliott: well it counted it as a winning entry of 20 bytes, so i must have done it right :P 04:18:59 I didn't really say all those things that I said. 04:19:02 oerjan: right :P 04:20:15 zzo38: O KAY 04:20:47 Some people said these kind of challenges are bad challenge. Actually they are just much more simple challenge than the other ones; it doesn't make it a bad challenge (as you can see from everything). 04:21:10 constant output is a bad challenge 04:21:11 always 04:21:14 constant, trivial output that is 04:21:17 as in short 04:21:25 because there are exactly two possible types of submission 04:21:34 - print it in the shortest normal way in the language (i.e. with the string embedded) 04:21:40 - exploit the interpreter/language to have the string built in 04:21:49 neither of these are interesting after, like, five billion such challenges 04:22:03 * pikhq mutters 04:22:04 invalid command name "::build::dc::flag_read" 04:22:06 Even the "Zero" challenge I came with a shorter solution in shell script that other people had, at first. And then someone else figure it out afterward. 04:22:19 elliott: ok but it was interesting the first time in unlambda, at least :P 04:22:20 pikhq: wat 04:22:28 elliott: Yes, I agree they are not as good as other challenges. But I don't call them "bad challenge". 04:22:32 I do 04:22:36 Ohwait. 04:22:42 because there haven't been many good challenges lately in anagolf at all 04:22:52 and the same person kept submitting bad challenges recently 04:22:54 ::groups is the namespace all my building semi-objects are in. 04:23:12 (I seem to have implemented half an object system without even thinking about it) 04:24:34 Can you do "Zero" challenge in five bytes of Bash or fish, or four bytes of Zsh, or five bytes of Zsh with symbol only? Maybe. It is not too difficult to figure out, although it is not the most obvious thing (it should be somewhat obvious once you figure it out, which is not too difficult to do) 04:25:58 Or, in Forth, it is very simple. But in symbol only, it is slightly more difficult? (I used a feature which I think is specific to gforth) 04:27:25 With the "PubSubHubbub" challenge, you probably cannot get any shorter than the obvious literal way in most programming languages, but some might be able to do shorter. 04:30:29 -!- augur has quit (Remote host closed the connection). 04:43:11 -!- pikhq has quit (Ping timeout: 252 seconds). 04:43:16 -!- pikhq has joined. 04:45:02 Do you know any Windows binaries of the C binding of LLVM? 04:45:16 :t readLn 04:45:17 forall a. (Read a) => IO a 04:46:07 Currently, I'm page 10 on Google's results for "gregor" 04:46:20 Some Mendel jerk is clogging up the works, but there are a few others too. 04:46:44 And this Samsa creature. 04:46:58 Gregor: well that must be a bug 04:47:27 *ba-dum* 04:47:37 why thank you 04:48:55 Argh, and of course some AOP guy beats me out if I look for gregor computer science 04:49:26 Heyoooo, #1 for gregor javascript though 04:54:12 The worst thing to be first for 04:54:22 Like coming first in a Terrible Person competition 04:56:23 -!- augur has joined. 04:57:33 -!- CakeProphet has joined. 04:57:34 -!- CakeProphet has quit (Changing host). 04:57:34 -!- CakeProphet has joined. 05:05:03 elliott: Thought you might like to know that I've got my stuff up on Github. 05:05:07 elliott: https://github.com/pikhq/town 05:07:03 -!- piz2 has quit (Ping timeout: 250 seconds). 05:07:24 Ohhh, so this is to tup sort of as autotools are to make 05:07:34 Yes. 05:07:51 What did you think I was doing? 05:08:01 Full stack 05:08:20 If I was doing a full stack, I'd have a reimplementation of tup going. :P 05:10:41 Something that I suspect is going to help me extensively when I get around to it is that my autoconf-alike and automake-alike are the *precise same thing*, thereby eliminating all need for a libtool-alike. 05:11:46 (as I think the *only* justification for libtool now is that Automake is too stupid to allow for outputting different rules based on what Autoconf gets.) 05:12:49 ... automake can output different rules based on autoconf. 05:13:15 Uh, no it can't. Automake runs before ./configure and potentially on a different system. 05:13:51 Automake generates a Makefile.in, not a Makefile, the rules that actually make it into the Makefile can be controlled by configuration. 05:13:57 It can conditionally *execute* rules, but that's not the same thing, now is it? 05:14:26 There are no "rules that actually make it into the Makefile"; ./configure just plugs in values for variables. 05:14:48 Dude, AM_CONDITIONAL. 05:16:03 That has got to have some *scary* implementation logic. 05:16:26 Okay, what's *really* going to help me is that my stack isn't in M4M4SHM4MakePerl. 05:16:52 And add another "SH" in there if you're using libtool. 05:17:14 How it works btw is that it prefixes every line with either @_TRUE@ or @_FALSE@, and I'll bet you can guess what those get set to :) 05:17:30 *Wow*. 05:17:40 That is brilliant. Utterly insane, but brilliant. 05:19:26 Another victory for not-M4Make. 05:19:27 :P 05:19:55 :P 05:20:18 I wonder if people would complain if I started writing all my C using M4 macros. 05:26:18 Almost certainly. 05:26:43 Not that that should stop you; the Bourne shell is written using CPP macros extensively, after all. 05:26:54 (to make it look like FORTRAN, IIRC) 05:28:55 Gregor: I don't particularly thing M4 is good for writing C programs but there is different opinions. It should be possible to invent programming language which M4 works very well as a preprocessor. 05:30:55 zzo38: M4 was designed to be a better C preprocessor. 05:31:42 Though it's pretty generic. 05:32:18 I however don't think it is particularly good for preprocessing C codes, in my opinion. 05:32:31 Oh, sorry, it was meant as a FORTRAN preprocessor. Never mind. 05:33:06 Maybe it is better for that; I don't know much about FORTRAN. 05:56:03 -!- Sgeo_ has changed nick to Sgeo. 06:04:42 -!- hagb4rd has quit (Ping timeout: 250 seconds). 06:09:25 How do I download the LLVM binaries (and header files) for using library with C? I read some things about it and have some idea to make something with it. 06:17:24 pikhq: back 06:17:41 zzo38: llvm has no binary builds, just compile it yourself 06:17:54 pikhq: mind if I offer critiques :D 06:18:01 elliott: Not at all. 06:18:13 pikhq: ok well first disclaimer i havent actually really read your code much at all 06:18:29 pikhq: in "c needs c99", the winning solver of c99 should take over the c command completely 06:18:40 why? because some compilers might use --link-with rather than -l, etc. 06:18:48 they need to know everything the program wants from the C compiler 06:18:52 and react in a way appropriate to them 06:18:59 now, this actually applies to all types of tools 06:19:11 and the c you "needs c99" on has nothing to do with the c99 that wins! 06:19:16 so I feel like it should actually read "needs c : c99" 06:19:25 i.e., I need the command c in my local namespace, that's a c99 06:19:32 elliott: Holy crap you're proposing it be more generic than Autotools. That's frightening. 06:19:43 and c99 would be aliased to {c c99} (because many solvers will want to find a C compiler, /then/ see if it supports that standard) 06:19:49 pikhq: perhaps so :) 06:19:55 elliott: But I would need to install all sorts of stuff to make it work and compile properly it seems, I want to download the binary 06:19:58 pikhq: but I think this model would actually allow it to work, unchanged, for Windows 06:20:01 zzo38: there is no binary 06:20:12 elliott: True, there is that. 06:20:25 pikhq: of course, most C compilers will just defer to the default functions to do things 06:20:30 The download page does list the binaries. But it doesn't seem to be the one I am looking for. 06:20:54 pikhq: the library-finder would return {link-with foo} rather than -lfoo, but it'd still have {cflags-ldflags foo bar} from pkg-config, there's no way to avoid that 06:20:59 and the compiler would just translate it accordingly 06:21:07 Is "LLVM Binaries for Mingw32/x86" the right one? 06:21:07 pikhq: the good thing is, I think this actually simplifies things 06:21:12 zzo38: oh, yes, it would be 06:21:20 pikhq: because right now the "c" command is kind of... weird and a pain 06:21:55 Do those binaries support the API though? 06:22:04 zzo38: download and see :) 06:22:05 pikhq: btw, does this generate the same Tupfile for a project no matter what machine you run it on? 06:22:07 OK 06:22:16 pikhq: because if so, I think that should be a separate step, like "town init" 06:22:25 and then town for the users will only generate the config file 06:22:28 No. It generates a Tupfile only on configuration. 06:22:30 which seems... cleaner to me 06:22:31 pikhq: I know 06:22:32 pikhq: but I mean 06:22:36 is the Tupfile always the same? 06:22:38 No. 06:22:40 I mean, it should be, since it should use @CC@ 06:22:43 and the like 06:22:44 pikhq: Why not? 06:22:54 If you don't use the tools correctly you will invent another autohell :) 06:22:55 Because, uh, I didn't make it do that. 06:23:10 Go make it do that X-D ...or I will 06:23:15 If I can figure out Practical Tcl Coding. 06:23:42 In addition to C99, you might also need to have one for C89, GNU89, GNU99, and the newer draft standards. 06:23:48 My Tcl is probably a bit hard to jump into; I tend to be a bit metaprogramming-crazy about it. 06:23:55 zzo38: Definitely do, it's just unimplemented. 06:24:05 zzo38: yeah, thankfully that can go in a separate library 06:24:29 (I imagine you'd also have libraries for things like "c needs qt" where listing all the libraries and cflags is a huge pain.) 06:24:55 elliott: Actually, Qt 4 does pkg-config, so it's not that big of a pain. 06:25:04 OK, but there's still horrible things :P 06:25:29 c++ needs libs QtGui 06:25:43 Bam, it pulls in Qt. 06:26:03 So yeah, what I am saying is: the "foo needs thing-that-actually-specifies-what-foo-is" thing is silly and I was silly for thinking of it. 06:26:41 pikhq: Have you got any up-to-date example files? 06:26:44 IIRC you did one for dc. 06:27:07 That dc one is what I've been using to test. 06:27:11 Got a link? 06:27:50 program dc {c needs c99;c needs libs gc;in-directory src/;c sources dc.c error.c main.c stack.c} 06:28:32 And such things as, SDL, X window system, POSIX, audio, libpng, zlib, cross-compiling to different targets, maybe even one for literate programming, etc. 06:28:52 pikhq: 06:28:53 program dc { 06:28:53 needs c99 06:28:53 in-directory src 06:28:53 c99 needs libs gc 06:28:54 c99 sources dc.c error.c main.c stack.c 06:28:56 } 06:29:07 ("needs c : c99" is stupid, it should just define it with the given name of the thing.) 06:29:21 pikhq: Ohbtw, does your gcc c99 finder still print out "Searching for c99" itself? 06:29:43 elliott: No, that's in the solve function ATM. 06:29:44 Because requirements (what I call things that have solvers for them) should instead just have a "friendly name" associated with them. (in this case, "a C99 compiler") 06:30:00 So in this case, c99 would be defined as an alias for {c c99}, which would have friendly name "a C99 compiler". 06:30:34 *Unfortunately*, it's also currently outputting that for cached results. 06:31:03 Can it cache to a file? It should be able to cache to a file. Specifically a file in ~/.cache. 06:31:12 Why? Because it could find a C99 compiler ONCE and never do it ever again for another project. 06:31:13 -!- azaq231 has quit (Quit: Leaving.). 06:31:21 This is what we in the biz call FINALLY CONFIGURE TAKES 0 SECONDS. 06:31:44 And then, maybe delete ~/.cache in case you need to change something, is it? 06:32:03 I'd have a "town flush" command to wipe it out or something. 06:33:04 elliott: Uh, the appropriate values for that can change rather *often* based on environment. 06:33:09 Like maybe a file in directory ~/.cache/towncache/* and then you can just delete it using "rm" command 06:33:23 And change the permissions in case you want to tell it not to use the cache. 06:33:24 Which is the reason that Autoconf never has enabled caching by default. 06:33:26 pikhq: How often will the appropriate C99 compiler change? 06:33:34 pikhq: How often will the appropriate commands to link to Qt change? 06:33:50 elliott: any idea why the pl plugin of lambdabot would spam a bunch of assembler messages when I try to cabal install lambdabot? 06:34:00 elliott: I do it fairly often — I tend to test things with multiple compilers. 06:34:26 CakeProphet: Show paste. 06:34:28 pikhq: Yeah, but that won't affect the AUTOMATIC finding. 06:34:41 pikhq: If you want to do that, you have to override it anyway, say with the CC variable, which should invoke special behaviour anyway. 06:34:50 I looked, in fact the LLVM binary file I downloaded seems to be the correct one, it has "include/llvm-c" and "lib/*.a" 06:35:09 elliott: not my paste, but it's the exact same problem: http://hpaste.org/46397/lambdabot_install_failure 06:35:17 Also, you seriously overestimate how long this will actually take to execute. 06:35:29 pikhq: I'm not, but caching is a good idea /anyway/. 06:35:51 pikhq: Be like tup, consider insane scaling; what if the entirety of Chromium was one big Town project? 06:36:05 CakeProphet: Ask #haskell? 06:36:18 hmmm, okay. it's pretty odd. 06:36:24 Yeah, it is. 06:38:13 Rather, I think you seriously misjudge how much time in ./configure comes from its tests and how much come from how it's often *too much shell to fit in cache*. 06:38:38 pikhq: I realise it will be practically instant, I'm just saying that if you can cache, you should, especially because it _will_ help in certain cases. 06:40:31 That would probably be a pain without doing a truly naive form of caching. 06:40:56 pikhq: Howso? 06:41:26 Namely, serialising all of the cached solvers to file, thereby replacing the entire cache, and hence needing a lock on the cache. 06:41:53 pikhq: Personally, I see the inability to cache as a design flaw... 06:41:57 But anyway, my other points are more important. 06:43:09 Also, caching can produce *wrong* builds. Consider the (retarded) case of LLVM (at least on Debian), where when the library gets upgraded, the actual *name of the .so* changes. 06:43:27 Making it so that the cached output of llvm-config is wrong. 06:43:39 Yeah yeah yeah, my other points are more important :P 06:49:04 I should also try and think of a better way of handling "Which group should I change the flags of ATM" for the solvers... 06:49:51 At the moment, each solver just looks at the namespace its caller is in. 06:50:09 pikhq: Eh? 06:50:12 Which, though functioning with how I've got it set up right now, is not very clean at all. 06:50:15 My method fixes all that. 06:50:24 A solver doesn't just set something; it literally returns a function. 06:51:12 Oh, duh, returning a function would help with that. 06:51:49 pikhq: Technically, it could return an integer; all "needs foo" has to do is do the solving magic for foo, and then bind foo to whatever the successful solver results in in the caller's environment. 06:52:18 elliott: I'm using the same solver scheme for libraries, though. 06:52:54 pikhq: And? 06:52:59 pikhq: The compiler handles libraries. 06:53:11 It can call out to a common library-finder that uses the solver routines if it so desires. 06:53:16 pikhq: In fact, this actually helps: 06:53:39 pikhq: A library solver that uses the technique "just link with an .so" would return the object (not function or set flags or anything) {link-with libname}. 06:53:48 Whereas a pkgconfig one may return {cflags-and-ldflags ... ...}. 06:54:03 The library-finder would then pass this back to the C compiler using it, which would translate it into settings of variables in its local namespace. 06:54:06 See? Simple and elegant. 06:54:35 That is absolutely elegant. 06:55:07 Lesson to be learned: MAKING THINGS FUNCTIONAL MAKES THEM BETTER ALWAYS 06:55:10 AAAAALWAYS 06:55:20 EVEN WHEN THE TASK IS "BE NON-FUNCTIONAL" 06:56:04 I don't think so. 06:56:37 o rly 06:56:53 Maybe. 06:57:56 Maybe not. 06:58:12 But I think I am sure. 06:59:23 pikhq: we can all learn from zzo38's wisdo 06:59:24 m 07:03:09 what are we talking about? 07:04:31 do you own that nick 07:04:37 i might have to steal it 07:05:21 I own it. 07:05:29 and have for years. 07:05:53 -!- monqy has quit (Quit: hello). 07:06:06 -!- PatashuPikachuRe has joined. 07:06:52 every time I sign on to freenode some punk has it and then has to be humiliated while I ghost and idenfity my nick back. 07:07:43 F: there's an option to make NickServ throw them off automatically, i think 07:08:11 oh? I'll have to check it out. 07:08:24 does anyone here know Go or Perl6? 07:08:32 -!- Patashu has quit (Ping timeout: 252 seconds). 07:09:32 i know some go, some perl6. 07:09:41 I thought all one-letter names were reserved for staffers. 07:14:21 That option exists, yeah, that's why I'm always guestified when I get d/c'd and autorejoin :P 07:19:17 Information from NS INFO doesn't match. 07:19:47 oh noes F is an evil liar 07:20:26 eek 07:21:41 * oerjan finds NS INFO inconclusive on that matter 07:21:50 o_o 07:21:52 r u sr 07:21:52 s 07:22:27 i mean it doesn't give evidence any way, since the nick isn't currently identified 07:22:39 are you actually doubting e 07:22:39 em 07:22:58 not really 07:25:40 I'll bet if I updated my reverse DNS hostname to libdl.so, people would think there was something broken in Freenode's software ... 07:26:09 no people would just point 07:26:10 and laugh 07:26:14 haha this noob doesnt have libc.so 07:26:15 what is he, poor 07:26:20 or, well 07:26:20 we would 07:26:31 :( 07:26:51 ur just not bourgeois enough for us Gregor 07:27:08 Gregor: I think your reverse DNS makes sense the way it is now, it doesn't need to be adjusted. 07:27:33