<<09-07-2012>>
00:52:16*SchalaZeal quit (Quit: Konversation terminated!)
01:31:21*Trixar_za is now known as Trix[a]r_za
05:02:24*SchalaZeal joined #nimrod
05:02:37SchalaZealDom, you on?
05:50:41*SchalaZeal quit ()
07:13:58*zahary_ quit (Read error: Connection reset by peer)
07:14:34*zahary_ joined #nimrod
09:31:50*SchalaZeal joined #nimrod
09:35:25*Trix[a]r_za is now known as Trixar_za
09:40:03SchalaZealThink tomorrow I'll bring Nimrod's PCRE wrapper up to date
09:40:33Trixar_zaThat should probably please dom96 considering Aporia uses it
09:40:34Trixar_za:P
09:40:41SchalaZeallooks like it's dependent on an old version
09:41:19SchalaZealanyway, forked, I'm off to bed
09:41:56*SchalaZeal quit (Quit: Konversation terminated!)
09:51:40zahary_hi Araq. I was not at home yesterday. sorry for not responding
09:56:38*zahary_ is now known as zahary
10:03:39*zahary left #nimrod (#nimrod)
10:05:25*zahary_ joined #nimrod
10:11:32fowli had to comment out a deprecated function in 8.14 but its already fixed in head
10:25:24*Trixar_za is now known as Trix[a]r_za
12:09:36*Araq_ joined #nimrod
12:10:46Araq_Amrykid: somebody actually tries to use your 'events' module :P
12:10:55Araq_please answer the guy on the forum
12:10:57Amrykid:o where
12:11:04Amrykidoh
12:11:43Araq_that's what you get for asking me to include in the stdlib :P
12:11:54Araq_oh and please rewrite it to use closures instead :-)
12:12:18Araq_that will be a good test for closures
12:12:59Araq_see you later, bye
12:13:01*Araq_ quit (Client Quit)
12:13:16Amrykiduh?
12:20:19Amrykidwhy does code-block not work on the forum?
12:21:52Amrykidnvm, the alternate code block syntax works
12:55:31*ccssnet quit (Read error: Connection reset by peer)
14:59:42*zahary_ quit ()
15:09:45*zahary joined #nimrod
16:33:12*SchalaZeal joined #nimrod
16:37:53Araqping zahary
16:38:03AraqI don't have much time but: why does
16:38:26Araqproc += [T: TOrdinal](x: var T, y: T)
16:38:31Araqdoes not work?
16:38:57Araqit works if I remove the 'distinct's in TOrdinal ...
16:48:13dom96hello
16:48:22Araqhi dom96
16:51:28dom96what's up?
16:51:43dom96btw the Pi ran out of memory.
16:52:06dom96That's why the bootstrap just failed D:
16:52:26dom96Is there anything I can do to tell nimrod to use less memory?
16:58:55Araqhow much ram do you have?
17:01:23dom96~200MB
17:02:11Araqthat should be enough by far ... bootstrapping takes 140MB on my machine ...
17:02:32dom96Yes, well gcc takes up memory too
17:02:38dom96so does nimbuild
17:02:40dom96and other apps
17:02:45AraqI see :-)
17:02:59Araqwhat about swapping?
17:03:05dom96In fact gcc takes up a lot of space
17:03:09dom96er, memory
17:04:50Araqwell you can do --genScript --compileOnly
17:04:59Araqand then execute GCC with the script
17:05:13Araqthen the Nimrod has exited and freed all the RAM
17:05:40dom96That could work.
17:05:47dom96Would require adjustments to the builder's code...
17:06:43Araqwell I dunno, maybe you shouldn't run nimbuild on your Pi ...
17:07:17AraqI will feel sad when I broke your SD by pushing too often ...
17:07:58Araqwe could also make Nimrod run the GC before executing GCC btw
17:08:08AraqI dunno how much that would gain though
17:08:34Araqit'd need to free some global data structures explicitely before that
17:08:44Araqas the GC has not much to do otherwise
17:09:31dom96I can just run it every so often.
17:09:45dom96That will lessen the pain on the SD card
17:11:47dom96But yeah, i'm trying not to use a swap file so that my SD card doesn't fry
17:14:37SchalaZealHey Dom, how would I go about making an AUR package for babel? Just install the resulting binary and leave it at that, or... is there a system directory I can install .babel files to?
17:15:47dom96SchalaZeal: Please don't make a package yet.
17:15:54dom96Wait until I release 0.1
17:16:01dom96Babel needs a large overhaul before it's ready
17:17:52SchalaZealah
17:18:05AraqSchalaZeal: about the cross compiling
17:18:10SchalaZealhm?
17:18:12Araqif targetOS == osWindows: exe = addFileExt(exe, "exe") # is in the code
17:18:25SchalaZealwhich file?
17:18:27AraqI suppose it should be 'if hostOS == osWindows'?
17:18:33Araqextccomp.nim
17:18:52SchalaZealah yeah
17:18:58SchalaZealthat'll work
17:19:16AraqI dunno ...
17:19:25Araqwhat if it does '--genScript'?
17:19:34Araqthen it should add the .exe for targetOS ...
17:19:45SchalaZealyeah
17:19:46AraqI suppose that's the proper fix then:
17:20:16dom9631 issues.
17:20:23Araqif (genScript and targetOS == osWindows) or (hostOs == osWindows) ...
17:20:25dom96Why do I always have a job whenever the issues pile up?
17:20:31SchalaZealIt's just having .exe applied to an ELF formatted executable just confuses Linux because it doesn't exist
17:20:38Araqsure
17:20:53Araqdom96: it's not like you can fix the issues :P
17:21:20SchalaZealI actually want to update the PCRE wrapper
17:21:40dom96Araq: That's cruel.
17:21:40AraqSchalaZeal: btw the code used to simply not add an .exe but that's *wrong* for windows
17:21:48dom96Araq: Now I'm sad :(
17:21:58Araqdom96: sorry, but it's true; most bugs there are hard :P
17:22:40dom96Araq: Fine. In that case I will just spend my time working and earn money instead of attempting to help with Nimrod :P
17:22:58Araqhey don't put words into my mouth
17:23:16Araqalright prove me wrong :P
17:23:23Araqand fix one
17:23:50dom96Hey, I already fixed a couple.
17:23:57dom96But apparently the way I fixed them is wrong...
17:24:07dom96And now I don't feel like refactoring my fix...
17:27:10AraqSchalaZeal: if you work on PCRE please provide a 'noDLL' option for windows
17:27:22Araqthe dependency is annoying
17:32:14Araqbrb
17:36:21SchalaZealif I work on the wrapper or the actual thing?
17:36:32SchalaZealoh wait
17:36:36SchalaZealnevermind
17:37:02Araqand check your PCRE version supports the JIT :-)
17:37:19SchalaZealI'm always up to date. Compiled 8.31 yesterday
17:37:29Araqgood
17:38:04SchalaZealwith as many features as possible enables -- 16-bit library, Unicode properties, JIT, etc
17:38:24Araqugh
17:38:30Araqno 16-bit please
17:38:40AraqNimrod doesn't support utf-16 anyway ...
17:38:45dom96bbl
17:39:02SchalaZealWell that's fine, it's a separate library
17:39:18Araqin fact utf-16 is kind of brain dead now that unicode contains more than 0xffff glyphs ...
17:39:40SchalaZealDon't the Japanese need it though?
17:39:57SchalaZealinternationally speaking that is
17:40:09Araqer ... read a bit about unicode
17:40:19Araqutf-8 supports every glyph too
17:40:31SchalaZealAlright.
17:42:10SchalaZealAs for a DLL-less dependency, I suppose I could do that but my only strategy of doing it would involve ugly code like the type Nimrod's posix wrapper has, explicitly importing from header files and then linking a static library
17:43:14Araqyeah well I merged all the .c files of PCRE once
17:43:17Araqand did:
17:43:26Araq{.compile: "pcre_all.c".}
17:43:45Araqno need for a static lib nor headers this way
17:43:45SchalaZeal............I guess that'll work.
17:44:12SchalaZealbut the source would have to be redistributed I think. Not that that's a problem with its license.
17:44:26Araqyeah that's what I did
17:44:33dom96back
17:46:16SchalaZealI think it'd be wise to have Nimrod interact with CMake or autotools in that case though, because explicitly setting up configuration would be cumnersome.
17:46:22SchalaZealcumbersome**
17:46:36AraqI only did it once
17:46:47Araqhardcoded the configuration I wanted it to have
17:46:54Araqit's not OS specific anyway
17:47:16Araqand even if it were, there is still the option to bind to the DLL
17:47:17SchalaZealalright
17:47:59Araqa native regex implementation in nimrod would be sweet too ;-)
17:48:11Araqbut lots of work and another uphill battle
17:48:32SchalaZealMy belief on static vs. DLL is a DLL preference because it's better to have a bunch of smaller dependencies depend on one module than several fat sized binaries hardcoding the same library
17:49:21Araqnah dynamic linking is stupid IMHO
17:49:55SchalaZealWell on embedded software, yeah.
17:50:10SchalaZealBut.... I find it useful otherwise.
17:50:25Araqthe OS could simply merge multiple definitions/pages in memory
17:50:42Araqand you don't get the dependency hell
17:51:01SchalaZealthat sounds taxing on memory though.
17:51:27Araqand the "only 1 place to fix" argument doesn't hold anyway; the bugfix could break compatibility
17:51:48SchalaZealWell okay, there's a good point.
17:52:03Araqplus most clients of the library most likely don't use the code that is infected by the bug ...
17:52:30Araqmany libs have lots of special functions but clients only use common stuff
17:52:45Araqwhich tends to not have many bugs left
17:52:59Araqoften bugfixes fix some weird corner cases
17:53:36Araqhurray, the corner case is fixed for every client; except that it didn't rely on this case in the first place ...
17:54:09SchalaZealheh
17:55:23Araqand as I said, the OS should deal with "wasted" memory so that static linking only takes up more harddisc space
17:55:59Araqwhich is only a concern on embedded systems
17:57:31SchalaZealWell I'll add a PCRE wrapper to my Nimrod fork later. Will make sure to allow a define of noDLL
17:57:48Araqexcellent
17:58:10Araqit will be compatible with the current pcre wrapper, right?
17:58:17Araqso that re.nim continues to work?
17:58:19SchalaZealI'll try
17:58:27SchalaZealhopefully so
17:59:07SchalaZealif defined(noDLL): import from header files else: use the DLL imports
17:59:40Araqthere is an easier way, checkout the {.pragma.} pragma :-)
18:00:16SchalaZealturns a procedure into a pragma if I remember, right?
18:00:29Araqwhen defined(noDLL):
18:01:13Araq {.pragma prceImport.}
18:01:15Araqelse:
18:01:32Araq {.pragma prceImport: dynlib: "pcre.dll", importc.}
18:01:32SchalaZealhmm...
18:01:44Araqproc x() {.pcreImport.}
18:01:54Araqit declares a user defined pragma
18:02:25SchalaZealand you just designate the proc as the definition?
18:02:31Araqyep
18:02:46SchalaZealAlright, sounds doable
18:02:53Araqcheckout opengl.nim for a full example
18:05:39SchalaZealhmm... importc without arguments imports a symbol with the same identifier as its Nimrod counterpart?
18:06:05Araqyes
18:06:24SchalaZeallooks like I have to adjust my own wrappers, hehe
18:08:12SchalaZealanyway, time to nap, see you in a few hours
18:08:16*SchalaZeal quit ()
18:20:53dom96Araq: hrm, maybe I could use tinyc on the RPI?
18:21:13Araqyou can try but I doubt it still works
18:21:19Araqtiny C is not maintained anymore
18:39:57dom96right. I might try it.
18:40:05dom96For now I need to go to the train station
18:40:06dom96bbl
18:40:11Araqbye
19:24:09dom96back
19:24:19dom96The train station was closed -.-
19:25:12Araqnimbot doesn't notice my pushes :-/
19:25:28Araqdo I need to trigger it manually again?
19:27:02Araqdom96?
19:27:24dom96sec
19:27:48dom96weird
19:28:05dom96Let me reboot and check whether maybe the github process crashed
19:28:39Araqwell I need to go
19:28:58dom96too bad
19:32:13Araqsee you later
19:32:34dom96bye
19:41:31dom96Araq: I'm not sure what happened. I'm assuming github is failing.
20:04:56*ccssnet joined #nimrod
20:45:58*SchalaZeal joined #nimrod
20:47:00SchalaZealWell I tried the cross compiler again with the latest revision, but this time, this happened: http://pastebin.com/T6absJ1P
20:49:38SchalaZealoh
20:49:44SchalaZealodd... nevermind
20:49:55SchalaZealgcc is the linker and not ld
20:49:57SchalaZealhmm
20:50:38SchalaZealwell it works when the linker is gcc and *not* ld
20:55:31dom96lol?
20:55:55dom96Maybe there is a simple mistake somewhere in the code
21:01:33SchalaZealNah
21:01:38SchalaZealit's a gcc thing
21:02:19dom96ahh
21:02:46SchalaZealI guess linkerexe is for compilers that have an inferior compiler
21:02:51SchalaZealerr
21:02:56SchalaZealinferior linker*
21:29:06AraqSchalaZeal: if you invoke 'ld' you need to provide all the static libs etc. that GCC passes to ld for you
21:29:18Araqso it's better to let GCC do the linking step
21:29:38Araqand yeah 'linkerexe' is for inferior compilers/linkers
21:31:59SchalaZealah
21:43:34Araqso dom96, should I pull the github trigger?
21:46:10SchalaZealhow do I merge Araq/Nimrod with my fork?
21:46:21SchalaZealerr into
21:46:48Araqdunno I keep forgetting the git commands that I never use ;-)
21:47:02SchalaZeallol
22:07:12*Trix[a]r_za is now known as Trixar_za
22:09:14dom96Araq: Yes, go ahead.
22:09:55Trixar_zaBtw dom96
22:10:18Trixar_zaI created an organization page, but it's not allowing me to upload
22:10:25Trixar_zaIt's annoying :/
22:10:51Araqyay at least the trigger worked
22:10:57Araqstupid github
22:10:59Trixar_zaerror: SSL certificate problem, verify that the CA cert is OK. Details:
22:11:00Trixar_zaerror:14090086:SSL routines:SSL3_GET_SERVER_CERTIFICATE:certificate verify failed while accessing https://github.com/SliTaz/Tazpanel.git/info/refs
22:11:00Trixar_zafatal: HTTP request failed
22:11:05*Trixar_za agrees with Araq
22:11:34dom96Trixar_za: Ask in #github>
22:11:36dom96*?
22:11:50Trixar_zaThey have a channel here?
22:12:34dom96yep
22:16:36Araqyay github works again ...
22:18:00dom96:)
22:20:51SchalaZeallooks like it's cd nimrod-git && makepkg -fi again...
22:21:49AraqSchalaZeal: forking is bad if you never make pull requests
22:23:13SchalaZealGit seems pretty complicated for me so... forking and pull requests....I know zilch about
22:25:32Araqgit and github are so common that everything is a simple google away
22:26:10SchalaZealhmm
22:27:35SchalaZealah that's what a pull request is
22:27:50SchalaZealhad it confused with fetch/merge upstream
22:32:43SchalaZealhuh... the fork count didn't go down when I got cold feet the first 2 times
22:32:53SchalaZealstill at 17
23:05:29*Trixar_za is now known as Trix[a]r_za
23:06:34AraqSchalaZeal: you're on windows, right?
23:06:48SchalaZealcurrently? no
23:06:49Araqoh never mind, you cross compile to windows ...
23:06:55SchalaZealbut I do dual boot
23:07:10Araqanyway can you have a look at http://forum.nimrod-code.org/t/55
23:07:15SchalaZealI have Nimrod on my Win32 partition as well
23:07:34Araqand see if you can figure out this weird stuff
23:09:12SchalaZealhttp://pastebin.com/nFnjBVSn
23:09:30Araqargh
23:12:30Araqshould now compile under windows
23:12:37AraqI have to sleep, good night
23:38:05SchalaZealoh wow.... hey Araq, remember how I told you I tried getting banned from Warcraft?
23:38:14SchalaZealerr wait nevermind