<< 15-09-2013 >>

---

01:25:00	*	EXetoC quit (Read error: No route to host)
01:34:49	*	Associat0r quit (Read error: Connection reset by peer)
01:44:55	*	Associat0r joined #nimrod
01:44:55	*	Associat0r quit (Changing host)
01:44:55	*	Associat0r joined #nimrod
02:30:42	*	q66 quit (Quit: Leaving)
02:34:28	*	Associ8or joined #nimrod
02:34:32	*	Associ8or quit (Changing host)
02:34:32	*	Associ8or joined #nimrod
02:36:31	*	Associat0r quit (Ping timeout: 264 seconds)
02:59:01	*	OrionPK quit (Quit: Leaving)
05:23:46	*	Associ8or quit (Quit: Associ8or)
06:02:16	*	brson quit (Quit: leaving)
06:36:06	*	MFlamer_ joined #nimrod
07:09:48	*	shodan45 joined #nimrod
07:12:55	*	MFlamer_ quit (Ping timeout: 250 seconds)
08:44:50	*	Associat0r joined #nimrod
09:18:15	*	Associ8or joined #nimrod
09:18:15	*	Associ8or quit (Changing host)
09:18:15	*	Associ8or joined #nimrod
09:20:18	*	Associat0r quit (Ping timeout: 264 seconds)
10:22:02	*	io2 joined #nimrod
10:33:37	*	q66 joined #nimrod
12:21:48	*	io2 quit ()
12:54:11	*	shafire joined #nimrod
12:54:13	shafire	hi
12:54:17	Araq	ho
12:56:42	shafire	how are you today?
12:56:56	Araq	busy
13:00:40	shafire	"As of version 0.8.5 the compiler is maintained in Nimrod. (The first versions have been implemented in Object Pascal.)"
13:00:45	shafire	Where there other languages too?
13:01:22	Araq	the first version*s*
13:01:39	Araq	the build system used to be in python
13:01:48	Araq	no other languages have been used
13:01:59	*	JStoker quit (Ping timeout: 246 seconds)
13:03:26	*	JStoker joined #nimrod
13:04:04	Araq	first I used the boehm GC but that kept crashing so that the pascal version simply leaked instead
13:09:18	shafire	ah okay
13:11:10	shafire	will be right back
13:11:25	*	shafire quit (Quit: Page closed)
13:40:47	*	shafire joined #nimrod
13:41:01	shafire	hi
14:21:25	Araq	ho
14:55:06	*	OrionPK joined #nimrod
15:03:00	*	ltbarcly quit (Ping timeout: 260 seconds)
15:36:15	*	guaqua` joined #nimrod
15:36:59	*	guaqua` is now known as guaqua
15:59:00	*	Associ8or quit (Quit: Associ8or)
16:29:14	Araq	hi guaqua, welcome
16:44:28	shafire	Araq:
16:44:40	shafire	why are there the {. .} statements?
16:44:53	shafire	Why do I need to say, that I am using a {. closure .}?
16:45:44	Araq	I wanted to support a mode that supports the common {} instead of indentation
16:46:06	Araq	also a mode that uses "end if" etc., hence "end" is a keyword
16:50:17	shafire	Araq: so the {} mode will come?
16:50:41	Araq	if somebody implements it
16:51:07	shafire	I don't know, if it is a good idea
16:51:38	shafire	Then, I have 3 options to choose
16:52:20	shafire	indentation, {}, (begin)?/end
16:58:56	shafire	See you tomorrow
16:59:52	Araq	no
16:59:59	Araq	you don't have to choose
17:00:19	Araq	you should use indentation based syntax because the alternatives suck
17:00:49	guaqua	using indentation saves you from extra syntax
17:00:54	guaqua	one should do indentation anyway
17:01:58	shafire	yeah
17:02:09	Araq	yeah but actually imo the tool should render the AST as you prefer it
17:02:36	Araq	I am tired of ascii text based tooling
17:03:16	Araq	so the idea was to support all 3 and this ensures a painless interoperability
17:03:41	Araq	which is also the reason for nimrod's style insensitivity btw
17:05:19	shafire	Ok, thanks for the explanation
17:06:59	*	shafire quit (Quit: Page closed)
20:18:07	*	Associat0r joined #nimrod
20:18:07	*	Associat0r quit (Changing host)
20:18:07	*	Associat0r joined #nimrod
20:53:29	*	BitPuffin joined #nimrod
21:30:44	BitPuffin	Araq: Binding bullet physics could be a cool (and useful) way to test out c++2nim
21:30:57	Araq	so do it
21:31:23	BitPuffin	Well, maybe
21:31:55	BitPuffin	I am not sure I would start using C++ libraries until modules can be separated into C and C++ stuff
21:32:12	BitPuffin	So that using a C++ libraries doesn't go all GPL on my code and forcing all of it to compile to C
21:32:15	BitPuffin	++
21:32:29	BitPuffin	Since the compiling to C++ is less mature right?
21:33:39	*	ltbarcly joined #nimrod
21:33:55	BitPuffin	Araq: do you get where I am coming from? :)
21:41:02	*	ltbarcly quit (Ping timeout: 264 seconds)
21:42:48	*	ltbarcly joined #nimrod
21:49:47	*	ltbarcly quit (Ping timeout: 260 seconds)
21:49:53	Araq	well bootrapping with c++ works
21:50:08	BitPuffin	It does?
21:50:13	BitPuffin	That's good
21:50:21	Araq	worked for me the last time I tried
21:50:24	BitPuffin	But how do we know it doesn't have a lot of hidden bugs
21:50:57	*	ltbarcly joined #nimrod
21:53:50	BitPuffin	Well
21:53:59	BitPuffin	isn't there a huge potential for regressions Araq?
21:55:26	*	ltbarcly quit (Ping timeout: 245 seconds)
21:56:00	Araq	when isn't?
21:56:39	BitPuffin	Well
21:56:43	BitPuffin	I mean compared to the C generator
21:56:53	BitPuffin	since that's what's been exposed the most
21:56:57	BitPuffin	slash tested
21:57:04	*	ltbarcly joined #nimrod
22:02:16	*	ltbarcly quit (Ping timeout: 264 seconds)
22:04:14	*	ltbarcly joined #nimrod
22:08:45	*	ltbarcly quit (Ping timeout: 245 seconds)
22:10:38	Araq	BitPuffin: the real problem is that some modules might not be compatible with c++ due to c++'s stricter type checking
22:10:53	Araq	I don't think you'll get any new crashes or runtime or something
22:11:10	Araq	when it compiles as C++ it works as C in my experience
22:11:11	BitPuffin	Araq: you mean in the stdlib?
22:11:19	*	ltbarcly joined #nimrod
22:11:22	Araq	yes
22:11:23	BitPuffin	hmm well then maybe it is worth doing
22:11:34	BitPuffin	still the isolation would definitely be nice to have
22:12:08	Araq	the isolation is like 5 lines in the compiler but I don't want to touch it until my talk is over :P
22:12:30	BitPuffin	although on the other hand I haven't kept up with the development of newton. Maybe that's gotten better these days
22:12:34	BitPuffin	Araq: Your talk? :)
22:13:22	Araq	yeah
22:14:13	BitPuffin	when are you doing a talk?
22:14:32	BitPuffin	tell me more this is exciting :)
22:15:51	*	ltbarcly quit (Ping timeout: 245 seconds)
22:16:40	Araq	https://thestrangeloop.com/preconf
22:18:29	BitPuffin	lol
22:18:32	BitPuffin	the J programming language
22:20:23	BitPuffin	Araq: Haha I thought the Clayton Bauman talk was nimrod related first :)
22:20:52	Araq	yeah. right.
22:21:16	BitPuffin	Araq: where is this conference?
22:21:28	BitPuffin	will it be recorded?
22:22:01	Araq	for the 1000th time
22:22:06	Araq	yes it will be recorded
22:22:12	BitPuffin	Yaaaaaaaaaaaaaaay!
22:22:22	Araq	and even if not, I can put the slides online myself
22:22:32	Araq	and have lots of notes written down to the slides
22:22:34	BitPuffin	yes but the important thing is that we hear you talk
22:22:40	BitPuffin	:D
22:22:44	Araq	:-/
22:22:58	BitPuffin	because then we can finally imagine your voice as we talk to you in irc
22:23:26	Araq	you could skype :P
22:23:38	BitPuffin	I suppose :)
22:23:50	BitPuffin	although first I have to fix my linux audio haha
22:24:27	Araq	but then you would know that I'm a single person and not a team of people :P
22:24:47	BitPuffin	haha
22:24:51	Araq	though I guess people only ever suspect Mike Pall is a team
22:24:56	BitPuffin	well, that never crossed my mind
22:25:04	Araq	yeah too bad
22:25:13	Araq	means I'm not productive enough :P
22:25:34	BitPuffin	but now you have me suspcious!
22:25:44	BitPuffin	hehe :)
22:25:47	BitPuffin	Well
22:25:58	BitPuffin	not only that you also have zahary to do some of your bidding :P
22:27:17	BitPuffin	I don't think anyone suspects that I am more than one person
22:27:21	BitPuffin	but I am gonna change that!
22:27:24	BitPuffin	it is my goal!
22:28:13	Araq	yeah well
22:28:16	Araq	I need to sleep now
22:28:23	BitPuffin	you do look very zen in that picture though Araq. It is like you are writing 1337 loc/s
22:28:29	BitPuffin	in your mind
22:28:46	BitPuffin	sleep well!
22:28:56	BitPuffin	if you can
22:29:08	Araq	well I do write lots of code sometimes
22:29:14	Araq	and never read it
22:29:27	BitPuffin	in your mind?
22:29:32	Araq	so the argument "code is read way more often than written" is false
22:29:34	BitPuffin	or for real
22:29:53	Araq	nobody *reads* thousands of lines of code
22:29:54	BitPuffin	Yeah I tend to write code more often than I read my own code
22:29:59	BitPuffin	I only read it when I change it
22:30:03	Araq	yep
22:30:12	Araq	it's only a concern if it doesn't work :P
22:30:18	Araq	anyway, good night
22:30:27	BitPuffin	but it is important to be able to read it when you will change it
22:30:29	BitPuffin	hehe
22:30:31	BitPuffin	night!
22:49:47	*	zahary quit (Quit: ~ Trillian Astra - www.trillian.im ~)
22:51:50	BitPuffin	#ifndef dLong
22:51:52	BitPuffin	#define dLong long long
22:51:54	BitPuffin	#endif
22:51:56	BitPuffin	too hard for c2nim to handle
22:52:30	BitPuffin	(52, 0) Error: identifier expected, but found '[NewLine]
23:17:45	Araq	BitPuffin: use #def instead of #define tehn
23:18:29	BitPuffin	Araq: shouldn't you be sleeping :)
23:18:36	BitPuffin	Thanks though I will try!
23:19:25	Araq	I can't sleep
23:19:41	BitPuffin	should all #define be #def Araq ?
23:19:57	Araq	you should read c2nim's documentation
23:20:21	Araq	it's sparse for a reason :P
23:20:50	Araq	the idea of the sparse docs is that people might read it this way ...
23:20:58	BitPuffin	http://nimrod-code.org/c2nim.html ?
23:21:37	Araq	yeh
23:22:14	BitPuffin	it seems like it should be able to handle define right there?
23:22:43	Araq	I don't think so
23:22:55	Araq	it's a type not a constant or template
23:23:09	Araq	c2nim wants to translate #define into const or template
23:23:15	BitPuffin	ah
23:23:22	BitPuffin	but where does this mention def?
23:23:41	BitPuffin	here we go
23:24:21	BitPuffin	so in other words pretty much any define can be a def?
23:24:41	Araq	yeah
23:24:56	Araq	unless you want to translate it :P
23:25:32	Araq	well anyway coming up with heurstics when it should handle it as a #def or a #define is a planned feature
23:27:03	BitPuffin	that would be nice
23:27:05	BitPuffin	hmm
23:27:22	BitPuffin	#else
23:27:24	BitPuffin	#def NEWTON_API
23:27:25	BitPuffin	#endif
23:27:28	BitPuffin	why does it whine about that
23:27:40	BitPuffin	same whine as before
23:28:12	BitPuffin	ah
23:28:16	BitPuffin	it wants it to have a type
23:28:25	BitPuffin	Will define work here Araq ?
23:28:58	Araq	nope #def looks correct
23:28:58	BitPuffin	nope
23:29:04	Araq	no idea why it does complain
23:29:47	BitPuffin	are there any known issues with nested ifdefs?
23:30:35	Araq	yeah well
23:30:58	Araq	c2nim needs to process these in a special way too
23:31:47	BitPuffin	I don't even know if this stuff is needed
23:31:59	BitPuffin	Basically it defines NEWTON_API
23:32:08	BitPuffin	which is like dllexport stuff
23:33:02	BitPuffin	Which basically means that these functions that use the NEWTON_API thingy in C before the function type (I don't know what it is called) Should use the importc pragma
23:33:19	BitPuffin	and perhaps dyimport etc if you want that
23:35:02	BitPuffin	So basically it handles weather the code is used as a static library or if it is a dynamic one
23:36:56	Araq	just do:
23:37:09	Araq	#def NEWTON_AP __declspec(dllimport)
23:37:23	Araq	and don't nest it in an #if(def)
23:37:30	Araq	this should work
23:37:36	BitPuffin	Araq: But what if it is a static lib?
23:38:22	Araq	--dynlibOverride, user defined pragma etc etc
23:39:10	BitPuffin	Now I am having trouble with an anonymous union
23:39:11	BitPuffin	or whatever it is
23:39:18	BitPuffin	it's a union { stuff }
23:39:23	BitPuffin	without an identifier
23:41:48	Araq	that's not C89 I think
23:42:15	BitPuffin	I don't have any idea
23:42:19	BitPuffin	but how should I deal with it?
23:42:25	BitPuffin	I can't think of anything
23:42:54	Araq	union { stuff } u; ?
23:43:02	BitPuffin	ah
23:43:05	Araq	or even better get rid of the union
23:43:05	BitPuffin	so that won't break the api?
23:43:23	BitPuffin	why would it even have the union
23:43:25	BitPuffin	I cannot grasp
23:43:31	Araq	and pick the most important field of the greatest size
23:43:47	BitPuffin	http://pastebin.com/7sbBnYwV
23:48:45	Araq	that's a nice union you got there
23:49:16	BitPuffin	the union is my friend
23:49:56	Araq	I especially appreciate the 'm_' and 'Newton' prefixes
23:50:28	BitPuffin	lol :)
23:50:28	BitPuffin	typical C though
23:51:03	BitPuffin	We can have prefixes stripped remember so that it is TBoxParam etc
23:51:25	BitPuffin	but I guess for that to be 100% nice we still need renamed imports hehe
23:51:43	BitPuffin	until then we are kind of stuck with the C way of prefixing it with library specific stuff
23:51:46	Araq	well the real question is why C has prefixes even though it is case sensitive :P
23:52:04	Araq	but I digress
23:52:34	Araq	well you should wrap the dFloat ... [64]; part of the union I think
23:53:20	Araq	with a bit of luck that's the maximum size of all the union's fields
23:55:55	BitPuffin	Araq: it depends on a __USE_DOUBLE_PRECISION__ constant
23:56:13	BitPuffin	or wait
23:56:16	BitPuffin	what do you mean wrap?
23:56:27	Araq	edit the C code
23:56:37	BitPuffin	Yes but how do you mean
23:56:39	Araq	and get rid of the union
23:56:47	Araq	and instead have
23:57:07	Araq	dFloat ... [64]; in the struct
23:58:10	BitPuffin	but what about the other stuff?
23:58:30	Araq	chances are you don't need to access it directly
23:58:48	Araq	and if you do you can provide accessor procs that 'cast' to the proper type
23:59:41	Araq	cast[ptr ProperType](addr(thestruct.dFloat))