<< 31-08-2013 >>
00:08:11NimBotAraq/Nimrod master f1d218c Araq [+0 ±5 -0]: fixes #569; C++ codegen works again
00:09:04Araqtime to get some sleep. good night
00:10:25BitPuffingood night!
00:20:09*Associat0r joined #nimrod
00:20:09*Associat0r quit (Changing host)
00:20:09*Associat0r joined #nimrod
00:29:11*ltbarcly joined #nimrod
00:30:32*ltbarcly quit (Client Quit)
00:32:01*ltbarcly joined #nimrod
00:41:37*DAddYE quit (Remote host closed the connection)
00:58:58EXetoCbeep
01:16:18*EXetoC quit (Quit: WeeChat 0.4.1)
01:24:36*darithorn quit (Ping timeout: 245 seconds)
02:01:58*q66 quit (Quit: Leaving)
02:26:27*darithorn joined #nimrod
02:43:28*DAddYE joined #nimrod
02:50:20*DAddYE quit (Ping timeout: 268 seconds)
03:29:10*OrionPK quit (Read error: Connection reset by peer)
03:29:24*DAddYE joined #nimrod
03:30:14*DAddYE quit (Remote host closed the connection)
03:43:21*ltbarcly_ joined #nimrod
03:48:09*ltbarcly_ quit (Ping timeout: 256 seconds)
03:53:31*ltbarcly_ joined #nimrod
03:54:36*darithorn quit (Ping timeout: 256 seconds)
03:58:10*ltbarcly_ quit (Ping timeout: 268 seconds)
04:31:11*DAddYE joined #nimrod
04:37:30*DAddYE quit (Ping timeout: 245 seconds)
04:39:01*wp_ joined #nimrod
04:40:05wp_yo, is it possible to set vhost (bindhost) when connecting to IRC using Nimrod
04:54:51wp_nevermind, I just hacked it in there in the ugliest way possible by explicitly binding to the hostname I wanted using irc.sock.bindAddr in irc.nim
05:06:20*wp_ quit (Quit: if you want to come bug me about bad programming practice, query WP on irc.coldfront.net)
05:34:12*DAddYE joined #nimrod
05:41:09*DAddYE quit (Ping timeout: 268 seconds)
05:53:59*ltbarcly_ joined #nimrod
05:58:29*ltbarcly_ quit (Ping timeout: 248 seconds)
06:04:11*ltbarcly_ joined #nimrod
06:08:40*brson quit (Quit: leaving)
06:08:45*ltbarcly_ quit (Ping timeout: 245 seconds)
06:13:01*Associat0r quit (Quit: Associat0r)
06:37:41*DAddYE joined #nimrod
06:44:10*DAddYE quit (Ping timeout: 245 seconds)
07:40:45*DAddYE joined #nimrod
07:47:17*DAddYE quit (Ping timeout: 256 seconds)
08:43:55*DAddYE joined #nimrod
08:51:06*DAddYE quit (Ping timeout: 264 seconds)
09:47:07*DAddYE joined #nimrod
09:48:09Araqhey DAddYE
09:49:51NimBotAraq/Nimrod master 78b35bc Araq [+0 ±2 -0]: fixes #563
09:49:51NimBotAraq/Nimrod master 9d98474 Araq [+0 ±1 -0]: fixes #459
09:49:51NimBotAraq/Nimrod master c592622 Araq [+1 ±2 -0]: fixes #554, fixes #179
09:53:47*DAddYE quit (Ping timeout: 246 seconds)
09:56:14*ltbarcly_ joined #nimrod
09:56:19*Associat0r joined #nimrod
09:56:19*Associat0r quit (Changing host)
09:56:19*Associat0r joined #nimrod
09:58:45*ltbarcly quit (Ping timeout: 245 seconds)
10:09:59*zahary1 joined #nimrod
10:10:58*zahary quit (Read error: Connection reset by peer)
10:15:51*ltbarcly joined #nimrod
10:20:26*ltbarcly quit (Ping timeout: 245 seconds)
10:26:12*ltbarcly joined #nimrod
10:31:07*ltbarcly quit (Ping timeout: 260 seconds)
10:37:03*q66 joined #nimrod
10:50:18*DAddYE joined #nimrod
10:56:40*DAddYE quit (Ping timeout: 245 seconds)
11:09:07*Associat0r quit (Quit: Associat0r)
11:49:50*EXetoC joined #nimrod
11:53:29*DAddYE joined #nimrod
11:54:07*EXetoC quit (Client Quit)
11:57:15*EXetoC joined #nimrod
11:59:43*DAddYE quit (Ping timeout: 245 seconds)
12:28:35EXetoCwos goin on, lads
12:38:14AraqEXetoC: as you can see I'm fixing bugs
12:39:37EXetoCoh, look at that
12:56:18*DAddYE joined #nimrod
13:02:28dom96hello
13:02:55*DAddYE quit (Ping timeout: 264 seconds)
13:04:28EXetoChi
13:14:02EXetoCok now it correctly detects the lack of discard in some cases, but now I can't return result by having "result" be the last expression; though that was a workaround that seems to have been fixed by one of the recent patches
13:14:53Araqthat was never planned to be supported
13:15:01Araq"result" as a last expression, I mean
13:15:17EXetoCok good
13:15:36Araqas it doesn't make too much sense and the (yet to be documented) rule is that usage of 'result' triggers a void context
13:16:40dom96Araq: Shall I check to see if your fix fixed the nimbuild corruption?
13:16:53EXetoCyeah
13:19:34Araqdom96: I don't think I fixed it :-(
13:19:45dom96:\
13:20:28Araqbut ok try it
13:22:01EXetoCok, this seems to work "proc `-=`*[T:float|float32|float64] (x: var T, y: type(x))", but not "proc `-=`* (x: var TReal, y: type(x))". should those mean different things?
13:23:11EXetoCmaybe I shouldn't mess about too much in system.nim, but I'm trying to figure out what shortcuts are possible
13:23:37AraqI don't think they should be different but let's ask zahary1, zahary_
13:25:13zahary1it should work, I'll take a look
13:25:51*zahary1 is now known as zahary
13:25:58Araqyou also broke tests/compile/tvarious (bug #544), zahary
13:25:58EXetoCsince when has the first approach been possible?
13:28:06zaharyI know, I really need to do one complicated change that will replace tyGenericParam with tyTypeDesc for the generic type parameters
13:28:41Araqwhy?
13:28:52zaharyI left some comments in the code about it
13:28:54AraqtyTypeDesc is quite different from tyGenericParam
13:29:10zaharytyGenericParam really means "any value" when it comes to sigmatch
13:29:26zaharynow that the generics also use sigmatch, this requires some ugly hacks there
13:29:41zahary"any value" in contrast to "any type"
13:29:48AraqtyGenericParam can have a list of constraints
13:30:25zaharyyes, I mean that the signature of the generic type should not use tyGenericParam
13:30:47zaharybecause it doesn't accept values, it accepts types
13:35:15Araqbut yeah, it means "any value" and not "any type"
13:59:33*DAddYE joined #nimrod
14:05:50*DAddYE quit (Ping timeout: 245 seconds)
14:23:14dom96Araq: So far it seems to be working.
14:23:45Araqyeah but it's always been "so far"
14:23:50dom96indeed.
14:29:24*Associat0r joined #nimrod
14:29:24*Associat0r quit (Changing host)
14:29:24*Associat0r joined #nimrod
14:33:33NimBotAraq/Nimrod master 4f7500d Grzegorz Adam Hankiewicz [+0 ±1 -0]: Fixes xmltree <> macro docstring.
14:33:33NimBotAraq/Nimrod master 810ed6a Araq [+0 ±1 -0]: Merge pull request #579 from gradha/pr_fixes_xmltree_macro_docstring... 2 more lines
14:37:24*ltbarcly joined #nimrod
14:41:48*ltbarcly quit (Ping timeout: 245 seconds)
14:47:43*ltbarcly joined #nimrod
14:52:13*ltbarcly quit (Ping timeout: 240 seconds)
14:53:42*OrionPK joined #nimrod
15:02:40*DAddYE joined #nimrod
15:09:10*DAddYE quit (Ping timeout: 245 seconds)
16:05:47*DAddYE joined #nimrod
16:07:26Araqhey DAddYE
16:09:44dom96Any university students in here? (Preferably from the UK)
16:12:42*DAddYE quit (Ping timeout: 264 seconds)
16:19:15profmakxdom96 will a postdoc at a university in the UK do too?
16:20:56dom96profmakx: Sure, that's awesome. What university (if you don't mind sharing)?
16:21:01OrionPKoo we got a scholar here
16:21:28profmakxdom96 st andrews
16:24:09profmakxOrionPK I only learned about nimrod at work
16:25:13dom96Am I blind or are the entry requirements not on this website? http://www.st-andrews.ac.uk/courses/route/USBECSCSCSC/year/2013-4
16:26:41profmakxdom96, if you want to do CS you should look at the CS site :P http://www.cs.st-andrews.ac.uk http://www.cs.st-andrews.ac.uk/prospective-ug/requirements
16:27:01profmakxhttp://www.cs.st-andrews.ac.uk/prospective-pg
16:27:16profmakx(depending on whether you are interested in undergrad or masters/phd
16:27:23dom96I see, thanks hah
16:27:42profmakxI think st andrews compsci is not quite as crazy as maths wrt a-level requirements
16:28:26profmakxadded bonus, we have one of the guys who perpetrated whitespace :P
16:29:07dom96hah
16:30:37dom96Interesting, Glasgow is 4th on the league tables for 2014.
16:32:00dom96profmakx: Can you tell me much about the course content for Comp Sci in St. Andrews?
16:34:37profmakxdom96, I did not do my undergrad in St Andrews, so I am probably not the right person to ask about details of the course
16:35:06profmakxI can tell you that we have a functional language group, a hci group, a CP group and a computer algebra group
16:35:09profmakxoh and systems
16:35:11profmakxwhatever that means
16:35:57profmakxI am myself involved mostly with computer algebra and maths, and know Edwin Brady and kevin hammond, since i am somewhat interested in dependent types
16:37:15dom96hrm, I'll certainly consider applying to St. Andrews then. What other universities would you suggest?
16:38:12profmakxfirst, what are you looking into? undergrad or postgrad?
16:38:23dom96undergrad
16:38:41profmakxand then, its probably good to choose by what you think your interests are at the moment
16:39:00profmakxst andrews is not only a good uni, its also a nice place to live. i don't particularly like glasgow :P
16:39:08profmakxcambrige is obviously good
16:39:11profmakxalso a bitch to get into
16:39:33profmakx(i don't like the city glasgow, the university has a good reputation)
16:39:46profmakxthen there's edinburgh
16:39:58dom96Yeah, I'm thinking of applying to Cambridge even though I doubt I will get into it. (Hoping that my experience will wow them)
16:40:12profmakxi have worked in manchester for a while, but haven't met any computer scientist there
16:41:21dom96The thing is, I will have to move no matter what university I choose in mainland Britain (I live in Northern Ireland)
16:41:42profmakxsure, if you can get in, cambridge is certainly a good place to be
16:41:44dom96And I've never actually been to any of the cities.
16:41:54profmakxcambridge and st andrews are very expensive in terms of living
16:42:00profmakxlike rent is absolutely crazy
16:42:23dom96oh :\
16:42:29profmakxplaces like edinburgh, glasgow, or manchester are a bit on the better side there
16:43:20profmakxin st andrews you are entitled to university accommodation as an undergrad
16:43:31profmakxbut i don't know how hard it is to get into the "cheaper" halls
16:46:08dom96what about Newcastle, is it any good?
16:47:00profmakxmhhm. I am not sure. I think they have a few good academics there, I don't know about teaching quality
16:47:08profmakxthe town is a bit on the manky side
16:48:49dom96hehe, I see.
16:49:45dom96thank you for your advice.
16:50:56profmakxno problem
16:51:08profmakxshould you come to st andrews for an open day or otherwise let me know
16:51:13profmakxi can show you around
16:51:45dom96Awesome. I'll let you know. Thanks.
16:54:31*profmakx goes back to deciphering arm architecture reference manual
16:59:40*enurlyx joined #nimrod
17:08:56*DAddYE joined #nimrod
17:15:00*DAddYE quit (Ping timeout: 245 seconds)
17:34:16*ltbarcly joined #nimrod
17:38:38*ltbarcly quit (Ping timeout: 241 seconds)
17:40:44*ltbarcly joined #nimrod
17:42:42Araqping EXetoC
17:45:20*ltbarcly quit (Ping timeout: 256 seconds)
17:46:45*ltbarcly joined #nimrod
17:51:22*ltbarcly quit (Ping timeout: 256 seconds)
17:52:16EXetoCAraq: hi
17:52:47*ltbarcly joined #nimrod
17:54:20AraqEXetoC: I can't reproduce bug #577, maybe because of the lattest changes to FP handling
17:57:19*ltbarcly quit (Ping timeout: 240 seconds)
17:58:18*ltbarcly joined #nimrod
17:58:20EXetoCI think the relevant code is similar now, and I don't get an error anymore when uncommenting the converter
18:02:23EXetoCweird sentence. the clamp is still there in my project is what I meant, and the converter doesn't cause an error, so yes it appears to have been fixed
18:03:03Araqstrange ...
18:03:06*ltbarcly quit (Ping timeout: 264 seconds)
18:04:03EXetoCthe fact that the FP changes might've resolved this?
18:04:17*ltbarcly joined #nimrod
18:04:35Araqgah, never mind
18:08:39*ltbarcly quit (Ping timeout: 240 seconds)
18:10:19*ltbarcly joined #nimrod
18:10:35EXetoCI confirmed your findings, so that's why I was confused by your response
18:11:47Araqwell I think I know the reason of the bug
18:12:08*DAddYE joined #nimrod
18:12:21Araqclamp is generic after all and you have a converter from float32 to circle
18:12:41Araqso both used to match with the same "fitness" value
18:13:57EXetoCok
18:17:38*ltbarcly quit (Ping timeout: 256 seconds)
18:18:17*ltbarcly joined #nimrod
18:18:45*DAddYE quit (Ping timeout: 245 seconds)
18:20:33*ltbarcly quit (Client Quit)
18:22:06Araqso ... how big needs the buffer to sprintf needs to be for %g, %f and %e ?
18:24:29*fowl joined #nimrod
18:37:54EXetoCWouldn't it be better to have float64 and float32 overloads in most cases? rather than float32 and float
18:38:11EXetoCalso, do we need UInt and Float magic? might be hard to change that though
18:39:45Araqfloat is used everywhere, float64 rarely, so I decided to do it this way
18:40:19EXetoCfair enough
18:47:14NimBotAraq/Nimrod master 7b7c68f Araq [+1 ±1 -0]: fixes #516
18:47:14NimBotAraq/Nimrod master 72678ab Araq [+0 ±1 -0]: fixes #398
18:47:14NimBotAraq/Nimrod master 18a5097 Araq [+0 ±1 -0]: Merge branch 'master' of github.com:Araq/Nimrod
18:55:16*darithorn joined #nimrod
19:01:47EXetoCserver.nim
19:01:50EXetoC[Warning] insufficient LD_LIBRARY_PATH
19:01:57EXetoCI don't think I've seen that before
19:02:22EXetoCmaybe I just missed it before. does it usually appear when running the tests?
19:09:31oalHas anyone used the JS compilation for any real world code?
19:10:44EXetoCThe official website, I think
19:11:34oalHmm, I don't see much JS in use there
19:14:06dom96EXetoC: Yes, it happens.
19:14:30reactormonkoal, define "real world". I had some fun with kwin.
19:14:48dom96The official website doesn't use the JS backend.
19:14:56oalSleep doesn't work when compiling to JS, as it's in the os module. setTimeout doesn't compile either. How would I call global js functions?
19:15:07reactormonkoal, import them yourself
19:15:11*EXetoC quit (Ping timeout: 260 seconds)
19:15:20*DAddYE joined #nimrod
19:15:27reactormonkand add them to the stdlib if you deem then useful
19:15:34dom96oal: You wrap them just like you would wrap C code.
19:15:38dom96Using the importc pragma
19:15:55reactormonkor importcpp if the JS part uses object syntax
19:16:04reactormonkaka foo.methodToWrap(bar)
19:16:07oaloh, they're defined in the dom package :)
19:21:40*DAddYE quit (Ping timeout: 245 seconds)
19:30:37*EXetoC joined #nimrod
19:42:44EXetoCAraq: The fact that there seems to be overloads for both 'int' and 'int64' is a mistake. it works though
19:43:03AraqEXetoC: what do you mean?
19:47:21EXetoCAraq: "proc abs*(x: int): int", "proc abs*(x: int64): int64" and so on
19:47:30Araqthat's not wrong
19:47:54*DAddYE joined #nimrod
19:47:54EXetoCit works, but do we even need the overloads for 'int'?
19:48:11Araqyes we do because 'int' is a first class type
19:51:32EXetoCbut isn't int implicitly convertible to either of them? where one has more priority than the other, depending on the platform
19:52:58oalIs there an example using the dom module, for compilation to js?
19:54:11EXetoCso, is the overload taking an int64 chosen when passed an int, on platforms where int basically is int64
19:54:39Araqthere is tests/js.nim, tests/js.html, oal for a minimum example
19:55:31AraqEXetoC: int+int64 -> int64, int+int -> int, int32+int -> int, int32 + int literal -> int32 etc.
19:57:15oalAraq, ok, thanks. How would I call window.setTimeout? It expects a cstring and an int. Trying window.setTimeout("", 1000) gives me "Error: value returned by statement has to be discarded"
19:57:56Araqwell discard it then or do not declare the return type or mark it as {.discardable.}
20:00:16oalWhy does it take a cstring and not a proc, by the way?
20:00:21*Mat2 joined #nimrod
20:00:23Mat2hello
20:00:38EXetoCAraq: ok, but why can't the 'int' overloads be removed, when both int32 and int64 overloads are provided? along with the other types in most cases.
20:01:26Araqbecause then the type depends on the target architecture, type(x+y) returns either int32 or int64 etc.
20:01:52reactormonkoal, strange.
20:01:54Araqseems like a bad idea to me, but maybe you're right and it would work out
20:03:06reactormonkoal, maybe connected to {.nimcall.}, but I'm not sure
20:06:04*fowl quit (Quit: Leaving)
20:07:21EXetoCAraq: I'm talking about the overload set of 'abs' for example, where it's a little strange to have overloads for int, in addition to int32, int64 and so on. but if the abs(int64) overload is somehow chosen when doing abs(int) on x64 example, then at least it's correct
20:08:17EXetoCconfusing, but correct nevertheless. I'm just trying to figure out if the "AbsI64" magic for example is ever referenced
20:08:49EXetoCor if it just works anyway, because of the size of intVal, if that's even relevant
20:09:23Araqyep. it's relevant and the reason it should work
20:10:00*adx joined #nimrod
20:10:12Araqhi adx welcome
20:10:18adxhello
20:10:19Mat2hi adx
20:12:41Mat2hmm, I have a const string: type caTest = "testA", a comparison like: if caTest != "test" is always false, why ?
20:13:08Mat2sorry const caTest = "testA"
20:14:26enurlyxHello, i have a problem with 'not nil' checks and the 'result' being a ref. I get the following error: cannot prove 'result' is not nil. I try to call new(result) and then a function on 'result' that expects the first param to be not nil. Is something wrong with this?
20:14:58Araqenurlyx: should work, bug report please
20:15:03AraqMat2: works for me
20:15:18Araqenurlyx: however
20:15:43enurlyxok
20:15:45Araqif your ref type contains not-nil fields, you need to use object construction syntax
20:16:06Araqnew(result) # ok, result not nil, but what about its fields?
20:18:29Araqdom96: please look at https://github.com/Araq/Nimrod/issues/491
20:18:34Mat2Araq: ok, found the error, wrong string content (vopy&paste fault)
20:25:21EXetoCAraq: ok, so it doesn't matter that AbsI is invoked when doing abs(int), no matter what the size of int? if so, what makes the 64-bit versions of the magics special?
20:26:03AraqEXetoC: I told you the 64 bit versions of the magics are obsolete
20:27:37EXetoCI'm not sure about that, but ok it's clear now. can those versions be removed?
20:28:13Araqsure
20:28:29Araqbut we have much better things to do
20:29:48dom96Araq: Why odes httpserver2 implement an http server when we have the httpserver module which does the same thing?
20:29:51dom96*does
20:30:24Araqit was some contribution
20:30:39Araqit supported more features than httpserver so I included it
20:31:04Araqmark it deprecated if it's not necessary anymore
20:31:25*Mat2 found out the ez80 word size is 24 bit.. nice.. great. *ARGH*
20:31:59AraqMat2: don't ask me to introduce int24 please ...
20:32:46*Araq awaits the day when somebody invents unsigned floating point numbers to save 1 bit
20:35:14Mat2Araq: no, if so you should better support the 20 bit extended word size for tms320x ;-), the second is already invented by the way
20:36:19Araqok sure; however ansi C doesn't support it yet. A real omission of course. Lets patch printf et al to deal with unsigned floating point numbers, in 32 and 64 bit variants
20:36:50Mat2*lol*
20:44:26enurlyxAraq:Yes, there are two fields that are nil. But same approach works if var is not 'result'. I make a testcase tomorrow. Good night.
20:44:50*enurlyx quit (Quit: Verlassend)
20:48:29*rubino123 joined #nimrod
21:22:22*brson joined #nimrod
21:32:03*ltbarcly joined #nimrod
21:36:59*ltbarcly quit (Ping timeout: 268 seconds)
21:42:22*ltbarcly joined #nimrod
21:43:11Araqping BitPuffin
21:47:02*ltbarcly quit (Ping timeout: 240 seconds)
22:01:25*Associat0r quit (Quit: Associat0r)
22:03:14*rubino123 quit (Quit: Leaving)
22:29:38dom96You guys heard of templeos? The language it's written in is rather interesting: http://www.templeos.org/Wb/Doc/HolyC.html
22:29:57OrionPKis that the one written by the crazy guy?
22:30:05dom96Yeah, he's schizophrenic.
22:30:15dom96from what i've read.
22:30:17OrionPKthen yes, I have heard of it :p
22:30:31OrionPKhe keeps renaming it, and he's a christian extremist schizo
22:31:09comex"* No type-checking
22:31:13comex"* No #define functions exist (I'm not a fan)
22:31:15comex"* No typedef, use class.
22:32:41Mat2dom96: Remarkly, I know a very similar OS which is based around a C IDE
22:37:34dom96Reading the main website... the first part of it sounds like random religious ramblings.
22:38:05Mat2probably a satire ?
22:40:02comexno, it's serious
22:41:04Mat2"It's like a modern, souped-up, multi-tasking,
22:41:04Mat2cross between DOS and a Commodore 64"
22:41:58EXetoCOrionPK: you should check out his square wave tunes on youtube
22:43:32OrionPKoh dear
22:43:44Mat2^that's a pattern encoded text. Its meaning is: between a souped-up cross, a modern DOS"
22:44:33Mat2with DOS = Denial-Of-Service
22:45:48Mat2I only see satirical content
22:45:57Mat2however, strange
22:49:03Mat2some seem to be hidden, political statements
22:50:17Mat2or the author is really just crazy, how knows
22:51:48EXetoChe is
22:52:07OrionPKyeah he's actually legit crazy afaik
22:52:12EXetoCor he really enjoys pretending, but probably not :>
22:54:14EXetoCand getting bashed like crazy
23:12:54*Associat0r joined #nimrod
23:12:54*Associat0r quit (Changing host)
23:12:54*Associat0r joined #nimrod
23:22:10*brson quit (Quit: leaving)
23:35:18Mat2ciao
23:35:21*Mat2 quit (Quit: Verlassend)