<< 28-07-2014 >>
00:10:58*bjz quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
00:29:05*ics joined #nimrod
00:30:37*Fx00F joined #nimrod
00:31:44adoniscikwhy can't nimrod automatically cast ints to cinsts as it does with cfloat/cdouble and floats?
00:32:10flaviubecause int != cint. int == size_t
00:33:24adoniscikso what's the right to feed (c)ints to C libraries?
00:33:32adoniscikmust I do cint(foo) for every int?
00:33:42adoniscikright way
00:34:06flaviuI personally like foo.cint, but that works too
00:34:20flaviuYou can also use a little documented feature, converters
00:34:53flaviuBut keep in mind that explicit is better than implicit
00:35:10adoniscikand what about arrays of cints; pls can you show you would define a simple array like [1,2,3,4,5] as cints
00:36:07flaviu[1.cint, 2.cint, 3.cint, 4.cint, 5.cint]
00:36:21adoniscikI don't need to tell you how ugly that is :)
00:36:33flaviuNot too nice, you might want to make a function to convert the array
00:37:18flaviuhttps://gist.github.com/aa0dda27a97eb89e03e0 would allow https://gist.github.com/a28bba00b3992edbc754 , which would look much nicer
00:39:23flaviubut you could just change the name of the proc in the second gist, same thing
00:42:15*bjz joined #nimrod
00:53:02adoniscikI'm trying to write a cint converter. I read the parameters using an int varargs but I couldn't figure out how to return them.
00:53:21adoniscikin the meantime I have another question: if int corresponds to size_t, what corresponds to cint?
00:55:37def-adoniscik: http://nimrod-lang.org/system.html#cint ?
00:56:35adoniscikI see, thanks.
01:01:49*ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
01:03:55adoniscikIs there a way to generate tuples without naming the fields? for example define converter `toCint`(t: varargs[int]): seq[cint] = map(t, proc (x: int) : cint = cint(x))
01:03:57*ics joined #nimrod
01:04:09adoniscikecho toCint(1,2,3) returns @[1,2,3]
01:04:27adoniscikcould I return a tuple do I can use the result of toCint for tuple unpacking?
01:09:53Varriountadoniscik: COuld you restate your problem?
01:11:08flaviuadoniscik: Keep in mind that cint's length isn't defined. It can be anything
01:11:16adoniscikhow can I elegantly convert an tuple or array of ints to cints ?
01:11:33flaviuFrom what I understand, old C compilers called int16 an int
01:11:44flaviuadoniscik: tuple?
01:12:07adoniscikyes, (int1, int2, ... , intn) -> (cint1, cint2, ... , cintn)
01:13:26flaviuHmm, that might be hard. Nimrod doesn't have variadic types
01:17:53adoniscikmaybe I dont need it. If I define a wrapper proc using int args and {.nodecl, importc.}, does nimrod automatically convert ints to cints?
01:18:07adoniscikI'm guessing not :)
01:18:36flaviuI don't know, but only one way to find out :)
01:23:04adonisciksure, I'm working on it. interestingly, you can do var x,y,z = (1,2,3) but not var x,y,z:int = (1,2,3)
01:23:52adoniscikif it had worked it might have been an elegant solution to my problem
01:27:17*bjz quit (Ping timeout: 255 seconds)
01:28:37*xenagi joined #nimrod
01:31:01*BlameStross left #nimrod (#nimrod)
01:39:23*saml_ joined #nimrod
01:39:33*ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
01:45:29*q66 quit (Quit: Leaving)
01:57:18NimBotAraq/Nimrod devel f345b02 Grzegorz Adam Hankiewicz [+0 ±1 -0]: Adds os.copyDirWithPermissions().
01:57:18NimBotAraq/Nimrod devel d597767 Grzegorz Adam Hankiewicz [+0 ±1 -0]: Adds hyperlinks to copyFile* docstrings.
01:57:18NimBotAraq/Nimrod devel 5c43fbc Varriount [+0 ±1 -0]: Merge pull request #1416 from gradha/pr_copy_dir_permissions... 2 more lines
01:57:38NimBotAraq/Nimrod devel b10bf62 Grzegorz Adam Hankiewicz [+0 ±2 -0]: Documents pretty command.
01:57:38NimBotAraq/Nimrod devel 83d7d30 Varriount [+0 ±2 -0]: Merge pull request #1412 from gradha/pr_documents_pretty... 2 more lines
02:10:59*darkf quit (Read error: Connection reset by peer)
02:12:02*darkf joined #nimrod
02:21:39NimBotAraq/Nimrod devel 90e3c99 Grzegorz Adam Hankiewicz [+0 ±1 -0]: Adds TSet.isValid().
02:21:39NimBotAraq/Nimrod devel cdecac1 Grzegorz Adam Hankiewicz [+0 ±1 -0]: Adds TOrderedSet.isValid().
02:21:39NimBotAraq/Nimrod devel dd47fa9 Grzegorz Adam Hankiewicz [+0 ±1 -0]: Moves TSet procs to their code block.
02:21:39NimBotAraq/Nimrod devel 6c3b967 Grzegorz Adam Hankiewicz [+0 ±1 -0]: Adds test cases for remaining TSet procs.
02:21:39NimBot6 more commits.
02:22:05adoniscikthanks for keeping us informed, nimbot :)
02:28:36*flaviu quit (Read error: Connection reset by peer)
02:30:31*flaviu joined #nimrod
02:44:47*Raynes quit (Ping timeout: 272 seconds)
02:45:00*boydgreenfield quit (Quit: boydgreenfield)
02:46:30*Raynes joined #nimrod
02:54:41*ics joined #nimrod
03:04:11*boydgreenfield joined #nimrod
03:15:52*flaviu quit (Ping timeout: 255 seconds)
03:18:22*boydgreenfield quit (Quit: boydgreenfield)
03:26:23*flaviu joined #nimrod
03:27:56*flaviu quit (Client Quit)
03:28:05*flaviu joined #nimrod
03:36:26*boydgreenfield joined #nimrod
03:45:59*xenagi quit (Quit: Leaving)
03:49:06*saml_ quit (Ping timeout: 260 seconds)
04:03:58*ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
04:38:34*ics joined #nimrod
04:43:37*mwbrown quit (Ping timeout: 256 seconds)
04:44:28*flaviu quit (Ping timeout: 240 seconds)
05:04:19*boydgreenfield quit (Quit: boydgreenfield)
05:04:32*ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
05:05:18*boydgreenfield joined #nimrod
05:05:22*Fx00F quit (Quit: leaving)
05:33:23*ARCADIVS joined #nimrod
05:40:00*ics joined #nimrod
05:40:24*boydgreenfield quit (Quit: boydgreenfield)
05:41:05*bjz joined #nimrod
05:41:40*jh32 quit (Ping timeout: 250 seconds)
06:01:10*Boscop quit (Read error: Connection reset by peer)
06:01:50*Boscop joined #nimrod
06:06:07*Boscop quit (Client Quit)
06:06:32*Boscop joined #nimrod
06:06:32*Boscop quit (Changing host)
06:06:32*Boscop joined #nimrod
06:11:12*bjz quit (Ping timeout: 260 seconds)
06:12:46*bjz joined #nimrod
06:33:21*krusipo joined #nimrod
07:25:42*kunev joined #nimrod
07:38:32*Araq_ joined #nimrod
07:38:55Araq_adoniscik: you can write [3.cint, 4, 5, 6]
07:39:07Araq_you only need to give the first element its type
07:41:58*bjz quit (Ping timeout: 260 seconds)
07:42:47adoniscikthanks @Araq_, I noticed that too.
07:43:09adoniscikis that trick mentioned in the manual? I use it all the time
07:43:19adoniscikwe should check
07:47:44*q66 joined #nimrod
07:51:31*bjz joined #nimrod
07:55:22*bjz quit (Quit: Textual IRC Client: www.textualapp.com)
07:56:10Araq_maybe, we should add it in any case
07:56:31*bjz joined #nimrod
07:59:23*io2 joined #nimrod
07:59:27*io2 quit (Changing host)
07:59:27*io2 joined #nimrod
08:02:06*bjz quit (Ping timeout: 272 seconds)
08:03:42*adoniscik quit (Ping timeout: 250 seconds)
08:04:20*bjz joined #nimrod
08:34:10*Araq_ quit (Quit: ChatZilla 0.9.90.1 [Firefox 29.0/20140421221237])
08:39:03*ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
08:50:52Skrylarhuh.
08:50:59Skrylartype mismatch, got (uint64,uint64)
08:51:05Skrylarfor the > operator
08:51:06def-Skrylar: import unsigned
08:51:11Skrylardef-: already did
08:54:07def-Skrylar: ah yeah, because they are no ordinals
08:54:28Skrylari had to remove the typeclass to get it to compile :/
08:54:35SkrylarI had a [T:int64|uint64] section
08:55:08def-Skrylar: sounds strange
08:55:31Skrylarvery
08:55:45Skrylarunsigned being a second class citizen is very inconvenient at times
08:55:53def-Yeah, I noticed the same
08:56:04Skrylarin this case, its serialization and packet code--a domain they are often used in
08:56:28def-I had problems with bigints, which store their parts in unsigned ints
08:58:08SkrylarI've been getting some really bad lisp pangs lately
09:05:47*bjz quit (Read error: Connection reset by peer)
09:05:47*bjz joined #nimrod
09:08:54*mal`` quit (Quit: ERC Version 5.3 (IRC client for Emacs))
09:14:42*mal`` joined #nimrod
09:27:46SkrylarOh my. Reading some old forth benchmarks out of curiosity
09:28:22Skrylarhttp://www.complang.tuwien.ac.at/forth/performance.html apparently the amount of source code is around ~25% as much needed for the C equivalents o.o
10:25:09*Kelet joined #nimrod
10:28:15KeletDoes anyone have a small graph implementation (nodes/edges)?
10:44:58*ARCADIVS quit (Quit: WeeChat 0.4.3)
11:01:29*AnnaJokela joined #nimrod
11:03:33*AnnaJokela quit (Client Quit)
11:19:03*Boscop quit (Read error: Connection reset by peer)
11:19:28*Boscop joined #nimrod
11:54:51*mwbrown joined #nimrod
12:01:09flyxhi folks! I'd like to expand the unicode package to add UTF16 & UTF-32 support. is anyone else working on something like this?
12:03:42flyxand can someone tell me what TRune16 is for?
12:12:10def-flyx: you mean toUTF16 and toUTF32 or what?
12:12:29def-ah, and the other direction too probably
12:12:34flyxdef-: exactly
12:13:06flyxdef-: I'm not entirely sure if it fits into the package, as only some functions have UTF8 in their name
12:13:24flyxI'd like to hear Araq's opinion on that
12:13:33def-yeah, since nimrod always implicitly assumes utf8
12:13:55flyxyes, but it *should* have support for other encodings
12:14:25flyxthey're required by some specs, like JSON and YAML
12:15:30def-and TRune16 is not used anywhere from what I can see
12:15:44flyxthat's why I wondered why it was there
12:24:06*flaviu joined #nimrod
12:31:51Boscophas anyone here heard of the Red language, and what do you think of it?
12:33:47*untitaker quit (Ping timeout: 272 seconds)
12:40:04*untitaker joined #nimrod
12:40:58*io2 quit (Quit: ...take irc away, what are you? genius, billionaire, playboy, philanthropist)
12:43:15*Araq_ joined #nimrod
12:43:41Araq_Boscop: I looked at Rebol once and it didn
12:43:45Araq_*didn
12:44:04Araq_*didn't convince me, but I forgot why ;-)
12:45:46Araq_flyx: feel free but maybe put it in a different module
12:46:26Araq_btw UTF16 and UTF32 suck ;-)
12:47:04flyxAraq_: I know. but it's needed in some places because specifications require it
12:47:18Araq_yup. specs suck
12:52:41*othello joined #nimrod
12:53:35flyxthe world will be infinitely better when software developers can always say „no we won't implement that, it sucks“ ^^
12:53:45othellohello, is there a way to check whether a type is a seq type at compile time?
12:54:03othelloe.g. type T = tuple[header: string, body: seq[int]]
12:54:35othelloI'd like to do sth like `when T.body is seq`
12:57:22Araq_doesn
12:57:36Araq_doesn't this work already?
12:58:07Araq_maybe you need: let dummy = x.body; when dummy is seq
12:58:54othellomy bad it does
12:58:59othelloNimrod is indeed awesome
12:59:28flaviuWell, if you want another way of doing it I came up with https://gist.github.com/fba6df9cb02bf10278d6 lol
12:59:29othellohowever `when T[1] is seq` doesn't compile
13:00:59flaviuothello: My overloading solution works there
13:02:03Araq_flyx: more realistically I think spec writers should be forced to implement their spec
13:02:23Araq_that doesn
13:02:37flaviuAraq: Remember how you said dom96's syntax would complexity the compiler?
13:02:37flaviuHere is the patch: https://gist.github.com/flaviut/aa0dda27a97eb89e03e0
13:02:45Araq_doesn't prevent bloat though
13:03:36Araq_flaviu: I'm not sure that's correct
13:04:10flaviuPossibly not, I was only concerned with getting it to compile. I doubt I'd be much more complex though.
13:04:50Araq_I
13:04:57Araq_oh ffs
13:05:03othelloflaviu: your solution works, but not at compile time
13:05:05Araq_stupid keyboard
13:05:30Araq_flaviu: I'd rather see real improvements in the compiler
13:05:42flaviuothello: works for me, https://gist.github.com/3970d93fbb16b3654b07 gives the expected results
13:05:58Araq_where "real" means "what I have in mind" :P
13:06:47flaviuWell, with that small change, you get type constructors, I think that's pretty real.
13:07:17Araq_so we have a "small change" that people soon will rely on
13:07:43Araq_that's worse :P
13:08:05*io2 joined #nimrod
13:08:55othelloflaviu: your gist works indeed, thanks!
13:09:27othelloflaviu: took me some time to get that you were building a MyT var with the `("foo", @[1,2])` bit ;)
13:10:05flaviuYep, all that MyT does is give fancy labels to the items in the tuple
13:11:59othelloflaviu: thanks again! Btw, is there a way to know the number of fields inside a tuple (at compile time I'd imagine)?
13:12:13*darkf quit (Quit: Leaving)
13:12:46Araq_othello: something in typetraits
13:13:09Araq_though it has some bug I don't remember if we fixed it
13:15:34othelloproc arity*(t: typedesc): int ?
13:15:45Araq_yup
13:15:53othellothanks!
13:17:53Araq_def-: iirc zmq is now a babel package, they changed the ABI
13:18:03Araq_so the old wrapper doesn't work anymore
13:18:23Araq_dunno why it's still in the main repo
13:18:48def-Araq_: ah, my bad
13:20:14def-Araq_: hm, can't find the babel package
13:20:36Araq_hmm strange
13:29:52*mwbrown quit (Ping timeout: 260 seconds)
13:33:08Araq_def-: maybe wrap nanomsg instead?
13:36:32*samlee joined #nimrod
13:36:42Araq_hi samlee welcome
13:37:00samleehello
13:37:06samleewhat?
13:37:08samleewhy am i samlee
13:37:31Araq_because of your mum
13:37:54*samlee quit (Remote host closed the connection)
13:38:11*saml joined #nimrod
13:39:00samli was setting up company irc stuff and had to use real name
13:39:04samlnow back to webscale name
13:48:37flyxirc, providing webscale names since before webscale was a thing
14:02:31flyxhm, just wondering, what's the best way of providing an exchangeable set of procs in Nimrod? to do something like:
14:02:40flyxlet encoding = guessEncoding(myString)
14:02:57flyxencoding.runeLen(myString)
14:03:10flyxencoding.encode(myRune)
14:03:11flyxetc
14:04:49Araq_you can either use objects + inheritance + methods or function pointer or 'case'
14:05:15Araq_I usually use
14:05:43Araq_case because the "oh this way somebody else can easily add an encoding without touching my code" never ever happens
14:07:22flyxwhat do you mean by 'case'? defining an encoding enum and provide that to every proc?
14:09:52Araq_yes
14:10:19Araq_you get a nice factory for free
14:10:28flyxwouldn't that be slower compared to function pointers?
14:11:45flyxhm, it's either a jump inside the function or one outside the function
14:12:20flyxprobably doesn't make much of a difference
14:12:41flyxI'll do that, thanks
14:15:10Araq_case is usually faster than function pointers
14:15:27Araq_in principle it always faster since the FP loses information
14:45:48*MayurYa joined #nimrod
14:45:48*MayurYa quit (Changing host)
14:45:48*MayurYa joined #nimrod
14:49:54*io2 quit (Quit: ...take irc away, what are you? genius, billionaire, playboy, philanthropist)
15:10:26*Araq_ quit (Quit: ChatZilla 0.9.90.1 [Firefox 29.0/20140421221237])
15:15:45*Demos joined #nimrod
15:28:49*kunev quit (Quit: leaving)
15:28:57Demosmustache braces are only used for pragmas right?
15:33:18Demosoh no they are used for sets as well
15:33:51def-Demos: and for key-value tables: {"foo": 12, "bar: 13}
15:33:54def-"bar"*
15:34:17def-and tables should be seqs of tuples*
15:36:54*Jehan_ joined #nimrod
15:42:11*othello quit (Quit: Page closed)
16:37:51*gsingh93 joined #nimrod
16:50:28*Matthias247 joined #nimrod
17:14:14*vendethiel joined #nimrod
17:17:28*ics joined #nimrod
17:22:39*Ven joined #nimrod
17:28:57*adoniscik joined #nimrod
17:30:03*brson joined #nimrod
17:30:09*brson quit (Client Quit)
17:30:12*ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
17:30:17*brson joined #nimrod
17:38:10*flaviu quit (Remote host closed the connection)
17:38:21*flaviu joined #nimrod
17:38:41*flaviu quit (Remote host closed the connection)
17:52:40*io2 joined #nimrod
18:01:06*Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
18:16:06*Ven joined #nimrod
18:17:41*ics joined #nimrod
18:21:38*flaviu joined #nimrod
19:10:28*Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
19:12:52*btiffin quit (Ping timeout: 250 seconds)
19:14:55*jh32 joined #nimrod
19:24:06*btiffin joined #nimrod
19:37:06*ics quit (Quit: My MacBook Pro has gone to sleep. ZZZzzz…)
19:41:56*Ven joined #nimrod
19:43:25*ics joined #nimrod
19:59:19*btiffin quit (Remote host closed the connection)
20:14:04*linkmonitor joined #nimrod
20:16:18dom96hi linkmonitor
20:17:57linkmonitorhi
20:23:58adoniscikhow's the link?
20:28:32*Ven quit (Read error: No route to host)
20:28:46linkmonitorReferring to me? If so, then well. If not, I apologize for logging into the middle of a conversation.
20:30:36adoniscikyes, just welcoming you to the club :)
20:32:55*Ven joined #nimrod
20:34:54Araqdon't worry there is no conversation going on
20:36:35Demosreally silent today actually
20:37:08adoniscikwhere's nimbot when you need him :)
20:38:31linkmonitorHa, well thank you much. I'm new to Nimrod, but it's the rising star in my list of favorite languages.
20:38:39linkmonitorHow's the adonis?
20:41:17adoniscikpretty sexy, according to mythology
20:42:01adoniscikand he's busy coding in nimrod :)
20:43:01adonisciknimrod is a mythological character too, so it's appropriate :)
20:43:54adoniscikis lambda a keyword?
20:44:28adoniscikyes, it is
20:44:48linkmonitorQuite appropriate. I never figured Greek mythological characters to be tech savvy, but there's is a realm unto itself.
20:45:42adoniscikis lambda reserved for anonymous procs or what? I couldn't find anything in the docs beside a brief reference to "lambda lifting".
20:46:28Demoswell lambda functions are anonymous procs
20:46:29adonisciksome closure business yet to be implemented?
20:46:48adoniscikbut anonymous procs already exist
20:46:55Demosclosures work as advertised, although I think there are some limits to having closures inside closures
20:46:58Demosyes
20:47:07adoniscikhttp://en.wikipedia.org/wiki/Lambda_lifting
20:47:11Demosoh lambda is a resurved word?
20:47:22Araqnah, Demos your knowledge is not up to date
20:47:28adoniscikyes, I just found out when porting a library that used it as an arg :)
20:47:30Araqthese limits are gone
20:47:42Demosyay!
20:47:55Demosand adoniscik try enclosing the name in ``s, like `lambda`
20:48:12Araqwell lambda is a keyword ... because ...
20:48:20Araqdo you really want to know?
20:48:36Demoskinda...
20:49:10*noam joined #nimrod
20:49:12Araqwell I figured how to do lambdas without lambda later
20:50:41Araqit's suprising how bad python is at multi-line lambdas when it's really not hard to do
20:55:35*Matthias247 quit (Read error: Connection reset by peer)
20:58:58*Johz joined #nimrod
21:00:52adoniscikhow would you defined a procedural type that returns foo and has no argument constraints?
21:01:40Demostype TProc = proc(): foo
21:05:50flaviuI'm trying to open a file, but its segfaulting
21:06:23flaviuthe path is correct, using the path in the terminal works fine
21:07:22flaviuOh, nm
21:07:41flaviuIt seems that the lines iterator takes a filename as a parameter, not a string
21:08:39Araqthere is also strutils.splitLines
21:09:30flaviulines is easier for me here, I'm reading a file anyways. Thanks though
21:10:07def-flaviu: i guess "type path = distinct string" would give a nicer error
21:11:48dom96Araq: You're going to love this: github redesigned its issues (again?) lol
21:12:04Araqoh god, really?
21:12:10Araqfor the better?
21:12:29dom96dunno
21:14:26*Ven quit (Quit: My MacBook has gone to sleep. ZZZzzz…)
21:21:31flaviudom96: This should really bug you: I'm parsing html with regexps :D
21:22:14Araqyeah it's not like we have some serious html parser in the stdlib already
21:23:21flaviuAraq: That example isn't exactly encoraging
21:23:26flaviuI can do it in two lines
21:25:47*betawaffle quit (Ping timeout: 264 seconds)
21:31:53*betawaffle joined #nimrod
21:43:33Araqso ... how do we let people override the new 'deepCopy'?
21:44:51NimBotAraq/Nimrod devel 69949c0 Grzegorz Adam Hankiewicz [+0 ±1 -0]: Escapes properly code blocks without highlite support.
21:44:51NimBotAraq/Nimrod devel 9b63a33 Andreas Rumpf [+0 ±1 -0]: Merge pull request #1423 from gradha/pr_escape_not_highlighted_languages... 2 more lines
21:47:17Jehan_Araq: In what context? Do you mean how to customize it for user-defined types?
21:47:38AraqJehan_: yeah
21:47:45Araqwell the major problem is syntax
21:48:48Jehan_Araq: What parameters does it get? source object, source heap, target heap, something like that? Or just a source object and the rest gets figured out automagically?
21:49:09Araqproc `=` (x: var T, y: T) # ok, obviously overriding built-in assignment
21:49:36Araqproc detroy(x: T) {.destructor.} # but this one needs a pragma
21:49:47*Johz quit (Ping timeout: 264 seconds)
21:50:09Araqproc deepCopy(x: T): T # and this one looks not special at all
21:51:31Jehan_So, you need to define all three procedures?
21:51:52Araqno, but the inconsistencies are weird
21:52:07Araqpeople voted for:
21:52:28Araqdestructor destroy(x: T)
21:52:48Araqbut then this doesn't improve much, if anything
21:53:19Jehan_Hmm.
21:53:54Jehan_What's the specific use case for the deep copy? Assignment with value semantics, replicating data for concurrency, etc.?
21:53:56Demoswell there is no way to resolve the inconsistency by changing destructors, since they do not show up as a symbol in code
21:54:26AraqDemos: yes there is, if we go for the "some proc names are special" route
21:54:48Demosoh true
21:55:20AraqJehan_: it's actually only invoked for inter thread communication
21:55:21Jehan_If you want special proc names, you can always do stuff like: proc `*destructor*`(x: var T) = …
21:55:27OrionPKwhy not free up the lambda keyword then
21:55:39AraqOrionPK: consider it done
21:55:59Jehan_Araq: Are source and target always in the same heap, and if not, how do you figure out which heap they belong to?
21:56:34AraqJehan_: well they are obviously not in the same heap ;-)
21:57:11Araqsource is some foreign heap, dest in its own heap
21:57:32Jehan_Araq: I expected not, but wasn't sure if you had some additional mechanism in mind.
21:57:53Araqit's a really simply mechanism really
21:57:54Jehan_Ah, so you figure out the source heap from the reference, and the target heap is whatever's current.
21:57:58Jehan_Gotcha.
21:58:04Araqyes
21:59:17Araqfor the first iteration only ref and ptr are allowed to be overriden for deepCopy
21:59:41Araqbtw I though about naming it threadTransfer() or something similar
22:00:07Araqbut we already have shallowCopy
22:00:17Jehan_Eiffel uses `deep_import` to distinguish it from copying.
22:00:42Jehan_However, import already has an established meaning in Nimrod.
22:02:25Araqwell I always was against __ASCII_ART__ but proc `*destroy*` is not too bad
22:03:21Jehan_Well, the idea is to have something that people don't accidentally use.
22:03:59Jehan_I wasn't suggesting the particular syntax, just something that forces you to be explicit about your intent.
22:05:06Jehan_I still think a pragma would look more idiomatic for Nimrod.
22:07:01Araqfyi we tried 'proc destroy' and it lead to weird bugs where some wrappers would already have picked that name
22:07:24Araqso yes, some way to be explicit about it is required
22:09:14Jehan_By the way, is `::` used for anything? I can't define a proc by that name.
22:15:50Araqyeah, operators with : are reserved
22:18:12Jehan_Gotcha.
22:36:45*io2 quit (Quit: ...take irc away, what are you? genius, billionaire, playboy, philanthropist)
22:39:21Araqwell what should the pragma be? .override?
23:04:47Jehan_override what?
23:05:07AraqdeepCopy and destroy
23:06:01Jehan_I'd simply use two different ones.
23:06:14Jehan_{.deepcopy.} and {.destructor.} or something like that.
23:09:11Araqmeh
23:14:54*darkf joined #nimrod
23:18:33*Jehan_ quit (Quit: Leaving)
23:19:27Araqgood night
23:19:34linkmonitorgood night
23:31:53*saml_ joined #nimrod
23:38:29*linkmonitor left #nimrod (#nimrod)