<<08-09-2012>>

01:23:46*q66 quit (Quit: Quit)
06:50:24*shevy quit (Ping timeout: 268 seconds)
07:08:55fowltrying to build nimrod from within a vm, if i do it from cmd i get errors about gcc not existing, if i do it from msys shell it tries to execute %CC%, %LINKER% etc
07:09:10fowla windows vm i should say
07:11:30Araqwell the .bat allows easy editiing
07:11:59AraqC% and %LINKER% are set in the .bat for you
07:12:10Araqbut if it doesn't work, adapt them
07:12:31fowli just had to put the mingw\bin on PATH >_>
07:12:40Araqoh and msys may not work
07:12:47Araqthe .bat is for "cmd.exe"
07:12:48fowlthe problem is msys emulates a unix shell
07:12:53fowlyea
07:13:04Araqwell yeah the .sh is not built for msys
07:23:09Araqwhat does 'uname -a' say on msys?
07:24:12fowlMINGW32_NT-6.1 HERMES-PC 1.0.17(0.48/3/2) 2011-04-24 23:39 i686 Msys
07:25:16fowlsetting all that up was tedious
07:26:01fowlnot sure where to get opengl from, the site said it comes with your video card driver, im pretty sure the guest additions didnt include it though
07:26:44*Boscop quit (Disconnected by services)
07:26:47*Boscop joined #nimrod
09:38:56*q66 joined #nimrod
10:46:02Araqhi q66
10:46:09q66hi
10:46:35Araqdo you feel like helping the bsd guy?
10:46:42q66didn't get to my VM yet
10:48:07Araqwell I'll answer him
10:48:14AraqI dunno what though :-)
10:48:43Araq"use a real OS :P "
10:49:11q66that's not an answer
10:49:14q66that's trolling attempt
10:49:29Araqoh freebsd 9 is actually quite new
10:49:42AraqI thought it was ancient
10:49:52q66yeah 9.0 is latest "stable"
10:49:58q669.1 is upcoming
10:55:33Araqwell the code uses sysctl to determine the number of CPUs
10:55:53Araqand apparently his sys/sysctl.h header does not compile
11:07:57Araqping zahary
11:11:28zaharypong Araq
11:19:32Araqbrb
11:42:20Araqwell, zahary, I've implemented arithop{|ops} * args
11:42:31Araqbut it's quite useless :-)
11:42:53Araqx + y * z - x # produces:
11:43:28Araq optOps([-, +, *], [x, y, z, x])
11:44:11Araqand now in the macro body you can try to get a binary tree back out of this ...
11:44:37Araqwhich is possible but mind puzzling
11:45:53Araqit's easier to just use callsite() and process the AST recursively
12:00:31Araqmaybe it should produce [x, y, z, *, +, x, -] ?
12:01:16Araq(Reverse Polish notation)
14:40:57*Trix[a]r_za is now known as Trixar_za
15:16:52Araqping zahary
20:04:04reactormonkAraq: anything else that would not require me to read an overly amount of code? :-)
20:04:16Araqhu?
20:04:21Araqhash for floats?
20:11:55reactormonkno thashes.nim?
20:12:38Araqit's tested implicitely via ttables I guess
20:18:52reactormonkdone
20:34:19Araqreactormonk: thanks but test is not good enough
20:34:31Araqthe compiler computes -0.0 at compile time
20:34:42Araqand could wrongly output it as +0.0
20:34:50Araq(and I think it does ... :P )
20:36:03Araqand here you go, no more unittest weirdnesses (I hope)
20:36:23Araqso your test should look like:
20:36:29Araqvar dummy = 0.0
20:36:50Araqcheck hash(dummy) == hash(-dummy)
20:49:09reactormonkoh
20:50:24reactormonkAraq: yep, +1 really helps.
20:50:54Araqyeah but I think it can trigger an INF :P
20:50:55reactormonkand yep, it doesn't fail with literals
20:51:11Araqyeah, outputting floating point is hard
20:52:09AraqI guess I can check whether 1.0 / +-0.0 is +INF or -INF to detect -0.0
20:53:49reactormonkhum
22:44:05reactormonkAraq: nice
22:45:33Araqnot my work ;-)
23:05:22*Trixar_za is now known as Trix[a]r_za
23:21:18fowlwhats the new nil statement? just `discard`?
23:22:18dom96yep
23:25:39*q66 quit (Quit: Quit)
23:32:52Araqit's really much more used in the compiler than I thought :-/
23:33:07Araqso I didn't activate the deprecation warning
23:33:55fowli still think you should have went with nihil or nuffin
23:34:10Araqno new keywords please
23:35:06Araqin fact, you can also use a comment instead: #nihil
23:35:19Araqbut I'm not sure I like that