<<31-05-2012>>

---

15:18:31	*	NimBot joined #nimrod
15:36:49	*	apriori_ joined #nimrod
15:58:20	*	Araq_bnc is now known as Araq
16:01:32	Araq	llm: parsing overhead of style insensitivity is neglible
16:02:23	Araq	and IMO it's a bit harder to do bad coding with it; nothing keeps you from USING_ALL_caAAAPS in cs languages either
16:02:59	Araq	except that you have to be consistent in your typos which is annoying if you're not the owner of the code
16:03:59	Araq	in fact, I chose between "cs with enforced naming scheme" and "si with a much less restricting naming scheme"
16:04:48	Araq	the enforced naming scheme never works in practice IMO because of abbrevs and identifiers coming from other sources
16:06:19	Araq	enforced naming scheme would also imply no '_' as I strongly prefer camelCase ... now who would have liked that? ;-)
16:07:07	llm	i've changed my naming schema over the last 17 years from everything to everything - i don't mind as long as its consitent
16:08:23	Araq	afaik all_underscores are easier to read -- there have psychological studies about it
16:08:29	Araq	unless you're trained!
16:08:33	Araq	then camelCase wins
16:10:24	llm	my primary work is reading through masses of code - all the while, and yes underscores are easier to read - in every language i need to touch :)
16:11:33	llm	but names like this seem to be not so uncommon in business code i_TheFile_write_length .... evil underscore, ungarian, camelcase style...
16:12:10	Araq	I personally can't stand all_under_scores, it's bad for my reading speed; I don't want to read the single words an identifier consists of in fact
16:12:39	Araq	I want to do proper lexing instead, I care about tokens, not words
16:13:32	llm	underscore helps weak developers building better names - maybe to long but better for understanding, camelcase seems to need to much brain energy
16:13:51	Araq	and this_thing looks too much like this thing
16:14:09	llm	underscores make the name just more technical, camelcase is design itself :)
16:14:41	Araq	so it hurts my reading speed quite a lot; it's especially bad in C code where this_style abounds
16:15:23	Araq	anyway my final argument for SI is that an editor can show the names as preferred since the compiler is not picky about it
16:15:48	Boscop	I prefer hyphen-separated-identifiers
16:17:35	Boscop	1. only one keystroke instead of two for underscore 2. it's not visually too low, underscore is nearly in the next line 3. it allows for acronyms in identifiers which camelcase doesn't really (e.g. SQLQuery looks weird)
16:18:36	llm	StructuredQueryLanguageQuery is the way to go :)
16:18:49	llm	ist like LCDDisplay :)
16:19:45	llm	do it the microsoft way SqlQuery
16:21:33	llm	in my delphi times it was hard to read all my lowered types alwasy with first char uppercase when used by some other developers...
16:22:52	Araq	hyphen-separated-identifiers don't fly for me either
16:23:06	Araq	I want to be able to write a[i-1]
16:23:13	Araq	not a[i - 1]
16:23:51	llm	for better readablity a [ i - 1 ]
16:24:12	llm	or even more spaces ;)
16:24:13	Araq	I prefer dense code over code full of whitespace
16:24:35	Araq	whitespace is only superficial nicer to look at
16:25:03	Araq	but doesn't help readablility as it has no meaning
16:25:28	Araq	well ok, you need a middle ground
16:25:37	Araq	can't get rid of all whitespace either
16:26:23	llm	for i prefer most is to the the scope of an variable on first look - members m_, parameters p_ ... , statics s_, globals g_ .... maybe that comes from years of refactoring legacy code
16:29:47	Araq	mMember and pParameter would work as well?
16:30:15	Araq	no need for underscores here IMO
16:31:53	llm	would work
16:32:06	Araq	what to do with renoX's feature request?
16:32:15	llm	but the underscore makes it more clear
16:32:33	Araq	ok ok, enough about it please
16:32:50	llm	renoX feature?
16:33:06	Araq	he wants --floatChecks:on per default
16:33:18	llm	everything is ok - as long as the language is expressive and powerfull, symbols are just names
16:34:57	Araq	I found the floating point checks more of an annoyance than a help; I really like +-INF :-)
16:35:13	Araq	I mean in FPC/delphi
16:35:23	Araq	so I didn't make it the default
16:35:29	llm	not per default - maybe you should combine some of those settings in an naughty-child-mode
16:35:51	Araq	you can already via pragma {.push.}
16:36:15	Araq	he argues the default should be 'on' for consistency with integer types
16:36:25	Araq	which also raise an exception on overflows
16:36:26	llm	ok - ever thought about partical release, debug-code?
16:36:52	Araq	dunno what you mean, you can have any mixture of compiler flags that you like
16:37:06	Araq	and you can override the default setting for certain sections of code
16:41:00	Araq	in fact, I'd love to have INF for integers too but the hw doesn't support it
16:41:14	Araq	saturated arithmetic can be very nice
16:45:04	Araq	instead we were given INT_MIN with no positive equivalent; ugly and bug prone ...
16:58:08	*	Boscop quit (Ping timeout: 240 seconds)
17:32:23	llm	i would prefer 3 build settings, 1. release, 2. debug, 3. naughty child (maximum checks) ... something like that
17:35:49	apriori_	rather call it pedantic :P
17:55:21	llm	perfect
18:06:52	Araq	'release' isnt even implemented in the compiler :-)
18:07:21	Araq	just edit your config to implement 'pedantic'
18:15:32	llm	make it more a standard - put something in the default config
18:18:08	Araq	you do it and make a pull request :P
18:30:01	llm	good idea - teamleader :)
18:30:17	llm	im in the south of germany - you?
18:32:09	Araq	same
18:33:28	Araq	extra points if you also update the docs to mention this switch ;-)
18:53:37	*	Boscop joined #nimrod
18:53:43	*	Boscop quit (Changing host)
18:53:43	*	Boscop joined #nimrod
18:54:38	llm	how many points do i need?
18:54:56	Araq	there are different types of points
18:55:06	Araq	rocker points
18:55:28	Araq	"hell of a guy" points
18:56:02	Araq	chick points
18:56:21	Araq	and you don't need them, you get them
18:56:44	dom96	And what can you do with these "points"?
18:56:47	llm	what is better rocker or chick points
18:57:11	Araq	chick points obviously :D
18:57:14	llm	buy fan articles under shop.nimrod-code.org
18:57:31	*	dom96 wants a nimrod T-shirt
18:57:40	dom96	I must have earned enough points for that over the years, right?
18:57:42	llm	you can get a "hell of a guy" t-short for some rocker points
18:58:15	Araq	10 rocker points are 1 "hell of a guy" point
18:58:24	Araq	and 10 "hell of a guy" points are 1 chick point
18:59:13	llm	what points do you use for yourself? one of the 3 official or special
19:01:08	Araq	bawcock points?
19:01:12	*	Araq looked that up
19:07:35	Araq	dom96: sure you owned a t-shirt
19:08:49	Araq	but we have no t-shirts yet
19:10:25	dom96	Hopefully some day... I want Nimrod mugs too. :P
19:10:37	dom96	The github ones are quite cool.
19:13:13	dom96	This is quite interesting: http://geeksta.net/geeklog/exploring-expressions-emotions-github-commit-messages/
19:15:42	Araq	yeah
19:17:55	Araq	most bugfixes for C++ code ...
19:21:02	llm	that makes VimL not better - lowest commits, highest anger .... :/
19:24:33	dom96	I find it hard to believe that PHP is second for expressions of joy.
19:25:02	apriori_	I find it hard to believe that this stuff is in any way representative..
19:25:15	apriori_	it actually can only measure "lack of control"
19:25:36	apriori_	I could yell hours and hours about some langauges - still this won
19:25:41	apriori_	't be visible in my commit messages
19:55:27	apriori_	how do I check for nil?
19:55:44	apriori_	e.g. for a PFileStream
19:56:07	Araq	if f != nil:
19:56:21	Araq	if not isNil(f)
19:56:44	apriori_	ty
19:56:57	Araq	if f.isnil.`not` # hm should try this one
20:14:17	Araq	ping zahary
20:16:20	zahary	pong
20:18:01	Araq	any progress?
20:22:03	zahary	I didn't managed to work the other night, but today before work did make some progress
20:22:56	zahary	I have to change somehow how the slurped files are remembered as the TPassContext is not available in evals.nim
20:23:18	apriori_	ohm, how do I append an element to a sequence?
20:23:31	Araq	apriori_: 'add'
20:23:38	apriori_	Araq: ok, thank you
20:23:52	zahary	my plan is to say that the .ast field in each module symbol is used to store auto-generated code that will be appended to the module (the needed include statements in this case)
20:25:28	Araq	why so complicated?
20:25:56	Araq	just have another implementation for mSlurp in evals.nim
20:26:44	zahary	the implementation is moved there, yes. problem is where do you remember the list of slurped files (I want to remember them in the module symbol)
20:27:08	Araq	no, not moving the implementation, have 2 implementations :-)
20:27:54	zahary	even this will have the same problem? the implementation is evals.nim won't have a way to add the slurped files to the current module
20:27:55	Araq	slurp(variable) does not need to affect the rod file's dependencies IMO
20:28:25	Araq	if we allow for IO in evals.nim we can't reasonably detect the dependencies anyway
20:28:45	zahary	why so? all of this is just needed to create more abstracted macros using slurp
20:28:46	zahary	var template = compileMarkDown
20:28:56	zahary	compileMarkDown "somefile.md"
20:30:56	Araq	and what about 'gorge'? store a hash of the .exe to determine whether it changed?
20:31:04	Araq	but that's not reliable
20:31:18	Araq	it could output a random number
20:35:08	zahary	I see 2 possible paths there. gorge takes a program and a standard input string
20:35:08	zahary	1) program timestamp/hash and input hash are used to always cache the output
20:35:09	zahary	2) we have the caching API we discussed once and it's completely up to the macro to define the caching behavior
20:35:37	zahary	what about names such as staticRead and staticExec instead of slurp and gorge?
20:35:43	Araq	I like 2)
20:36:41	*	Bosc0p joined #nimrod
20:37:23	*	apriori_ quit (Ping timeout: 260 seconds)
20:38:00	zahary	I like 2) better too, but would you be so kind to implement the caching API for me :)
20:38:17	Araq	alright
20:38:48	*	Boscop quit (Ping timeout: 245 seconds)
20:39:03	zahary	what about staticExec?
20:40:39	Araq	meh, can't I ever have fun ...
20:40:58	zahary	heheh
20:41:03	llm	what your talking about?
20:41:38	Araq	llm: compile-time readFile and execProcess and their implications
20:41:49	Araq	for the module caching
20:43:00	dom96	When you say caching, is it caching for .rod files?
20:43:06	Araq	dom96: yep
20:43:28	dom96	So how would that work with applications that change their output based on files on the filesystem?
20:43:44	Araq	it wouldn't :P
20:43:49	dom96	Like say for example I want to get the current git commit hash of the app i'm building using gorge... :P
20:44:02	llm	getting better and better - .rod files remind me of good old pascal .tpu or delphis .dcu
20:44:33	Araq	llm: yeah but it was MUCH harder to implement for nimrod :-)
20:45:09	llm	pascal and delphi still are very easy in its features-set and interfacing problems
20:45:44	llm	the world is trivial without real templates, generics, macros...
20:45:56	Araq	indeed
20:47:11	*	apriori_ joined #nimrod
20:47:18	dom96	Araq: So I won't be able to do that?
20:47:33	Araq	dom96: the plan is to have some API so you program the caching mechanism for the advanced stuff
20:47:54	apriori_	Araq: btw.. I don't think its a good idea to allow to shadow the implicit result variable in a proc
20:47:57	apriori_	I just made that mistake :P
20:48:00	dom96	Araq: Ahh, that's good.
20:48:14	Araq	apriori_: yeah common newbie mistake :P
20:48:19	apriori_	hehe
20:48:32	Araq	can't really avoid it as the shallowing is essential for templates
20:48:47	apriori_	ok
20:48:50	Araq	but the compiler could produce a warning
20:49:07	apriori_	it could whether the implicit result variable was touched at all
20:49:14	apriori_	when using that implicit reutrn of it
20:49:18	apriori_	*it could check
20:49:24	Araq	excellent idea
20:50:04	Araq	and 2 lines of code :-)
20:50:14	apriori_	good :)
20:50:33	apriori_	the error message could contain a hint for that common mistake
20:50:48	Araq	it will only be a warning
20:50:53	apriori_	ohm, yeah
20:51:05	apriori_	although..
20:51:14	apriori_	when would it make sense to not touch the result value?
20:51:29	apriori_	ah, well. just allow it...
20:51:37	Araq	I bet I have code that does just that :P
20:51:40	apriori_	some templates could rely on that
20:51:44	Araq	for whatever reasons
20:52:03	apriori_	yeah, the simplest case is a default to nil fallback for a template
20:52:09	apriori_	or.. "default to init"
20:53:10	apriori_	didn't wrote much yet.. but so far I like nimrod...
20:53:17	apriori_	especially when I read about that "marshal" stuff...
20:53:45	apriori_	an issue which is entirely ignored by e.g. D... there are mediocre library solutions though
20:54:52	*	filwit joined #nimrod
20:55:10	filwit	hi guys
20:55:25	dom96	hello filwit
20:57:03	filwit	I am writing a feature to Aporia where pressing 'return' twice with back-tab automatically (unless you hold shift) like in Word Processors... but i just found out that GTKSourceView automatically goes to the beginning of the line when shift-return is pressed...
20:57:20	filwit	damn you, GTKSourceView!
20:57:46	filwit	*will back-tab**
20:58:23	filwit	now I have to start over. :V
20:59:12	dom96	What are you going to implement instead?
21:00:01	filwit	same thing, I just have to do it differently. I got it all working, but it was having odd error and I didn't know why until I tried Shift-return in GEdit
21:00:58	filwit	the idea is to have an "automatic indentation" option in settings, to help navigate in and out of proc/body indentation
21:01:09	llm	do nimrod suffer from differences between release, debug, what-ever mode when templates are used in library interfaces?
21:01:19	dom96	filwit: There already is one.
21:01:28	dom96	filwit: However, it uses gtksourceview's stuff.
21:01:40	filwit	what do you mean?
21:02:07	dom96	All it does is it indents when pressing enter if the last line is also indented.
21:02:36	filwit	ahh, that's not what I'm talking about
21:02:48	Araq	llm: it tracks the mode
21:03:05	Araq	when you change from debug to release it's as if the module has been touched
21:03:12	filwit	see, i want it to dedent by two spaces (or the defined tab amount) when you press enter and are already on a blank line
21:03:35	llm	so no mixing of debug and release version possible
21:04:00	dom96	filwit: Yeah, I know what you mean. But you should reuse the same option in the settings for it I think.
21:04:19	filwit	dom96: and when you press up/down, the cursor is snaped to the 'end' (only when lines are blank)
21:04:43	dom96	filwit: I'm not sure what you mean by that.
21:05:01	filwit	well let say you have:
21:05:05	filwit	proc foo =
21:05:09	filwit	bar()
21:05:16	filwit
21:05:22	Araq	llm: it's the best one can do, think of "when defined(release): ..."
21:05:25	filwit	# < cursor here
21:05:40	filwit	there is one "blank" line after "bar()"
21:05:54	filwit	that has two spaces in it
21:06:19	filwit	so when you press up-arrow, it moves the cursor to the end of the line
21:06:35	filwit	so it moves the cursor back into proc foo's space
21:07:11	Araq	filwit: do you know delphi's editor?
21:07:17	filwit	which works in conjunction with the auto dedent thing for navigating through functions
21:07:20	filwit	Araq: no
21:07:25	filwit	never used it
21:07:34	Araq	it has the nice property that the cursor is free to navigate everywhere
21:07:59	Araq	it just moves up as if the above line was long enough
21:08:06	dom96	filwit: hrm, i'm not sure I like that. You can just press the left-arrow in this case and it will do what you want.
21:08:13	llm	i've worked in a large c++ project and the main problem was that huge parts of the code were doing costly math, and some modules needs to be debugged - in the end there were a need to refactor all libs down to c interfaces to use some libs as debug others as release build - my hope is that nimrod will define something like an debug<->release module abi
21:09:02	dom96	filwit: It seems pointless to have two keys which do the same thing. Especially when you might want to go directly up, not have it indented.
21:09:22	filwit	dom96: what if the cursor is on a line that already has spaces? you'd have to press left multiple times. The point is to move in-n-out of the block bodies automatically
21:09:45	filwit	but not everyone is going to want that, and that's why there needs to be another option for it if it's there at all
21:10:11	filwit	dom96: if you want to go directly up you press shift-up
21:10:12	llm	thats shows very fast the less-good-for-prototyping features of C/C++, something like that is for example in pascal or delphi nearly not possible - due to its featureset - debug or release seems to be both very fast in this environments
21:11:20	dom96	filwit: Pressing shift should cause the text to be selected.
21:11:20	filwit	dom96: my idea comes from word processors (writing lists), basically when you're writing a list item, you press return and it makes a new list, you press return again and if the list item is empty it "exits" the list all together and you can get back to writing paragraphs
21:11:39	filwit	dom96, you're right I forgot about that
21:12:13	Araq	llm: in the worst case you need to refactor the nimrod code into DLLs I suppose
21:12:18	filwit	dom96, I guess there won't be any shift-up thing, it will just automatically indent to the previous block
21:12:24	dom96	filwit: Sure. Implement it, I might like it when I test it out :)
21:12:37	dom96	brb
21:12:56	llm	Araq, this was for an cutting/abrasion 3d simulation ... - because of this i asked yesterday if its possible to mix debug, release-mode in modules or libraries
21:13:24	llm	so parts of the code are checked,debugg-able - and others parts are not
21:13:57	Araq	well it is possible, but for now you need to edit the code and litter it with push and pop pragmas
21:14:07	filwit	dom96, the other thing I'm going to try and make tonight is commenting out things, what is your preference on hot-keys?
21:14:28	filwit	dom96, block-comments**
21:14:50	Araq	though module specific settings per configuration file are not hard to do
21:15:00	Araq	*to implement I mean
21:15:02	filwit	dom96, I'm use to Visual Studios Ctrl-K-C/U, but I'll make it whatever you prefer
21:16:54	Araq	so it's "staticRead and staticExec" vs "slurp and gorge". opinions?
21:16:57	llm	Araq, how does "in the worst case you need to refactor the nimrod code into DLLs I suppose" fits with "well it is possible, but for now you need to edit the code and litter it with push and pop pragmas" - is it or not?
21:17:14	*	apriori_ quit (Ping timeout: 245 seconds)
21:17:43	Araq	llm: well these are the 2 solutions that I see
21:18:35	llm	fine, then nimrod supports debug, release mode library and in-module debug/release mixing - wow thats great
21:24:24	*	Bosc0p is now known as Boscop
21:24:25	*	Boscop quit (Changing host)
21:24:25	*	Boscop joined #nimrod
21:26:10	filwit	Araq: Ya I like staticRead/Exec because they're more descriptive and probably not wildly used
21:26:17	llm	that is a great feature if you try to debug an 100 times called function - whichs needs 10mio calls of some different functions - which produces so much content (voxel-like data) that it is no way of logging the data to get faster to the debug-point
21:27:51	filwit	Araq: that way you're not guessing what's going on with the code, or needing to look up something. If you see staticExec("myprogram") anyone's going to know what it does, but gorge("myprogram")? idk..
21:33:24	dom96	filwit: I don't have a preference so go ahead with the Visual Studio way. I bet Araq is used to that as well.
21:33:37	filwit	k
21:36:33	filwit	what does the 'end' keyword do btw? I can't find it in the index
21:36:49	dom96	It's probably just reserved with no purpose.
21:37:05	dom96	No purpose currently at least
21:37:08	filwit	ah, okay
21:37:23	Araq	not true
21:37:30	Araq	it's used in the templating system
21:37:42	filwit	okay
21:38:04	dom96	Araq: What is it used for?
21:38:11	Araq	'end if'?
21:38:17	dom96	oh that.
21:38:17	*	apriori_ joined #nimrod
21:38:42	filwit	I thought it might be there and do nothing just to make the Ruby folks feel comfortable :)
21:39:23	Araq	in fact, my initial design was to allow for indentation, {} and 'end if' syntaxes
21:39:42	Araq	so that the IDE can show the AST in the style you prefer
21:39:51	filwit	ya I know Ruby has a "variable" syntax style
21:40:06	filwit	what changed you mind?
21:40:21	filwit	you just found better uses for those things?
21:40:31	Araq	nothing really, I started with the indendation based syntax as this is the most restrictive to design
21:41:08	Araq	and now I think I'll use () instead of {} because the expr/stmt distinction is about to be removed anyway
21:41:32	Araq	the pragmas are kind of {.ugly.} for this reason
21:41:43	filwit	interesting
21:41:45	Araq	so that {} is available
21:42:00	dom96	something tells me some people will not like this.
21:42:23	Araq	(and of course I would never use anything else than the indentation based syntax anyway)
21:42:42	filwit	i'm sure having {} available would increase popularity. I'm fine with any style so long as it works, but I've met some people that wont give a second look unless it looks like C to some degree
21:43:12	Araq	filwit: but these are people I don't care about ;-)
21:43:23	filwit	^ true :)
21:43:28	ponce	hi, I work in c++ maintenance and wouldn't care forced indentation (if it's the debate)
21:45:15	Araq	ponce: kind of, yes ;-)
21:46:24	ponce	could be a good thing since mixed spaces/tabs tend to creep in
21:47:46	filwit	well it would be interesting if the syntax actually supported multiple styles, much like how Nim's case-insensitivity and "_" allows for any combo of camelCase, PascalCase, or classic_case, but it's nothing I'm to concerned with myself
21:48:35	filwit	then again, i could see it actually confusing people
21:49:22	Araq	ponce: the compiler forbids tabs
21:49:28	ponce	great
21:49:36	filwit	but, you never know, having a language where you can basically "write like you want to" might attract a lot of folks. and Like you said, maybe the IDE could actually parse and display each style based on a user preference
21:49:57	filwit	anyways, random pointless musings are pointless :)
21:50:18	dom96	The IDE should first have a proper suggest feature :P
21:50:35	filwit	yep
21:50:41	Araq	yeah, dom96, I try to fix it this weekend
21:50:56	dom96	Araq: Great. Will you make it run as a service then?
21:51:05	filwit	is the compiler designed to stay open and accept AST updates "in memory"?
21:51:09	Araq	ugh ...
21:51:18	Araq	filwit: that's what dom96 wants
21:51:28	Araq	but it requires some changes to the compiler
21:52:03	Araq	the "skinnable" syntax has a huge disadvantage when it comes to documentation though
21:52:22	filwit	well if that where the case, it would mean that any syntax changes made in the compiler would automatically reflect in the IDE suggest
21:52:23	Araq	we can already see it for microsoft's .NET documentation
21:52:38	filwit	I'm not sure what a "skinnable" syntax is :S
21:52:40	Araq	"select your language"
21:52:49	filwit	oh I see
21:53:05	dom96	We could just make the docgen translate into the different "skins" :P
21:53:06	filwit	yes, it's benefit is debatable
21:53:17	*	llm quit (Quit: ChatZilla 0.9.88.2 [Firefox 12.0/20120420145725])
21:53:34	filwit	and not really important at this point
21:53:54	dom96	of course. Many more important things left to be done.
21:56:44	Araq	what? Urls allow for \\ instead of // ?
21:57:13	dom96	what?
21:57:22	filwit	only on browsers I think
21:57:41	filwit	but I could be wrong about that
21:57:50	dom96	Araq: What do you mean?
21:58:35	Araq	http:// # <-- can be \\ too?
21:59:08	Araq	nah, firefox doesn't support it
21:59:24	filwit	dom96: do you know if I can cast[gunichar](char)?
21:59:52	filwit	without consequence?
22:00:54	dom96	That seems like it would cause issues.
22:01:05	dom96	gunichar is a unicode char.
22:01:11	Araq	filwit: what's wrong with a type conversion?
22:01:12	dom96	I'm not sure how it's represented in gtk though
22:01:25	Araq	it's an int32
22:01:50	filwit	hmm.. how do I type convert it?
22:01:56	filwit	char(gunichar)?
22:01:57	Araq	if it's an ascii char, you can just do gunichar(ch)
22:03:10	dom96	filwit: Are you trying to fix https://github.com/nimrod-code/Aporia/issues/4 by any chance?
22:03:22	dom96	(If not then could you try?)
22:03:46	filwit	i'm not, but I can get to it
22:04:21	filwit	I'm trying to get my intelligent-tabs to work
22:05:04	dom96	I fear it might require quite a bit of work. Gedit provides a way to select the encoding, would be nice to have that too.
22:05:27	filwit	ya I spoke before I looked :)
22:05:48	filwit	but I'll try and see what I can do if it's a critical issue
22:06:19	dom96	Brilliant, thanks :)
22:06:40	Araq	nimrod's encoding module may be easier to use than the gtk stuff
22:06:48	Araq	but lacks an encoding guesser
22:06:59	filwit	we have a lot of features that need to be done.. jump to error, source outline, better suggest
22:07:14	*	apriori_ quit (Ping timeout: 245 seconds)
22:07:15	Araq	yeah! jump to error ftw!
22:07:31	filwit	unfortunately we can't get it to even scroll on load LOL
22:07:43	dom96	Go to line works :)
22:07:50	dom96	Jump to error should be easy.
22:07:59	filwit	it does? why aren't we using that on load?
22:08:18	filwit	we can just get the line the cursor's at and jump to it then
22:08:26	dom96	Well, I tried using the same code of course.
22:08:44	dom96	I think the problem is that the sourceview widget is not initialized when the scrolling happens on load.
22:09:02	filwit	ya that would explain it's erratic behaviour
22:09:09	filwit	behavior*
22:09:25	dom96	As far as I know there is no signal which gets fired when the widget is fully initialized
22:09:41	Araq	os.sleep(10)?
22:10:05	dom96	Araq: That wouldn't be very reliable.
22:10:24	Araq	it's a quite common "idiom" for UI code ...
22:10:25	dom96	Well first of all we would have to use gtk's timers.
22:10:31	filwit	dom96, is there a "fully initialized" flag?
22:10:44	filwit	cause we could just sleep and ping them for completion
22:11:04	dom96	filwit: Hrm. I'm not aware of such a flag, but maybe it does exist.
22:11:24	dom96	I think i've actually tried using timers at one point and it did work.
22:11:36	filwit	k. well I'm going to try and get this tab thing done (cause it's more fun ;) but i'll help with these things after
22:11:44	*	apriori_ joined #nimrod
22:11:57	dom96	filwit: It's a real pity we're in a different timezone.
22:12:04	filwit	ya
22:12:15	filwit	you could always stay up to wee hours of the night
22:12:16	dom96	I should be able to stay up later tomorrow though :)
22:12:40	dom96	I need my sleep for tomorrow though, got an exam.
22:12:56	filwit	okay, get rest for that
22:13:15	filwit	but tomorrow I should be available to work with you
22:13:29	dom96	great
22:27:01	Araq	good night
22:27:21	filwit	bye
22:28:52	dom96	Yeah, me too.
22:28:59	dom96	See you tomorrow filwit
22:29:01	dom96	good night
22:29:06	filwit	bye as well
22:40:51	apriori_	how do I get a pointer to a value?
22:41:15	apriori_	ah, addr
22:41:21	filwit	if it's unmanaged you use: addr(obj)
22:52:39	*	Bosc0p joined #nimrod
22:52:50	*	apriori_ quit (Read error: Connection reset by peer)
22:52:59	*	Boscop quit (Disconnected by services)
22:53:06	*	apriori_ joined #nimrod
22:53:15	*	Bosc0p is now known as Boscop
22:53:16	*	Boscop quit (Changing host)
22:53:17	*	Boscop joined #nimrod
22:54:25	filwit	hi Boscop
22:54:30	filwit	from the D forums I see
22:54:43	filwit	me too
22:58:39	Boscop	yeah
22:59:46	*	apriori_ quit (Read error: Operation timed out)
23:12:31	*	apriori_ joined #nimrod
23:14:58	Boscop	filwit, I haven't posted much on the forums though
23:15:37	Boscop	most active in the channel, but not recently... busy with other stuff
23:16:38	*	apriori_ quit (Ping timeout: 244 seconds)
23:18:23	*	apriori_ joined #nimrod
23:50:28	*	apriori_ quit (Ping timeout: 260 seconds)