« first day (1903 days earlier)      last day (2945 days later) » 

3:01 AM
@AlexA. Oh his question was worded a little weird at first
I thought he meant he had 137 but it only said 127
 
¯\_(ツ)_/¯
I want to open a bass guitar store just so I can use the slogan "All Your Bass Are Belong to Us"
 
Then shouldn't it be for rentals only?
 
@AlexA. I'll bet that store exists already.
 
Probably :(
Well there go my hopes and dreams
 
3:07 AM
@QPaysTaxes I already suggested a few hundred ideas, so... ¯\_(ツ)_/¯ :P
 
@Geobits :(
 
@QPaysTaxes Well, you can at least take the concepts of the grouped ones.
 
Let's keep that clean
 
What do you mean by clean?
 
3:15 AM
If you want a clean wall of text, don't tell people about it so no one will write on it.
 
Hey guys, I'd be interested to hear about bugs you've recently ran into
 
@NathanMerrill I had the flu last week... pretty shit
 
@NathanMerrill i was just sitting on my desk and there was this huge pincherbug... I squashed it pretty quickly though...
 
the mantra of my new language is "The ideal language is the one you never write bugs in", so I'm hoping to include language patterns that (hopefully) mitigate them
-.-
 
@Downgoat You didn't eat it? I thought goats eat anything.
 
3:19 AM
@DrGreenEggsandHamDJ ಠ_ಠ
 
squashing is like cooking for goats
 
only tin cans, grass, everything, and noodles
 
it just prepares the meal
 
@NathanMerrill smh, we also roll it in mud...
 
seasoning?
nevermind, misunderstood
 
3:21 AM
I hard deleted a row from a DB and then tried to look it up two lines later an hour ago
 
@NathanMerrill One that I recently squashed was that I forgot to define an attribute in a class and the problem showed up quite a bit later.
 
@El'endiaStarman that's easy to catch with static typing (I'm guessing your language wasn't statically typed)
 
@NathanMerrill yeah
 
static typing and throwing exceptions on type mismatches kills most common bugs
 
No, static typing wouldn't have helped. Do you mean like making a variable always be one type?
 
3:23 AM
@AlexA. Drink alone is catching up.
 
well, static typing also means that you can't access an attribute on a class that didn't declare the attribute
it throws a runtime error
 
The attribute was defined by the super class and wasn't overridden by the child class.
 
You should drink alone while you try to learn C++
 
oooh, that is a different issue
 
Basically, I forgot to override a default. Not sure there's anything you can do in terms of language design to fix that.
 
3:25 AM
@Quill Turn off any ide checking, and take a shot for each syntax error.
3
 
perhaps I can.
can you explain what the default was?
or is that too complicated?
 
@NathanMerrill I feel like you ask nice questions in here and expect good responses and usually the responses are jokes. :P (I can't think of a bug I've encountered recently tbh)
 
@Geobits Well, I know how to spend my next few weekends
 
@AlexA. yeah, but I joke with them, so I can't point fingers
 
@AlexA. yeah, sorry @NathanMerrill :(
 
3:26 AM
@Quill I'm not accusing. I did the same thing last time he asked a question.
 
@Quill In the hospital? :P
 
@NathanMerrill ;)
 
@Geobits :D
 
@NathanMerrill Well, my Pytek parser converts AST_node objects (like AST_number) into pObject ones (like pInteger), and the parser knows which pObject types to try and convert to by looking at the objType attribute that the AST_node has, which is an empty list by default.
 
@Geobits Challenge mode: Write the code in Microsoft Word with size 8 Times New Roman font and 4" margins.
 
3:28 AM
@El'endiaStarman does an empty list ever make sense, or would each subclass feasibly override that attribute?
 
@AlexA. comic sans
 
by "I forgot to override" he really means "quartata forgot to override and I had to fix it"
 
@Quill I should have said Arial. Have you ever tried to write code in Arial? It's literally the worst.
 
Forgot to specify the pDirective
 
@NathanMerrill Maybe I don't want an AST_node to be converted because it won't be included in the object tree.
 
3:29 AM
Hey @AlexA. I just noticed that the tag is still around. Should we start closing/locking the questions with it and edit the wiki to reflect the status of the tag?
 
@quartata Hahaha, but I have encountered that problem before, I think.
 
@Geobits And drink alone is now the leader. I guess I'll get some whiskey then.
3
 
so, I think I know how I'd solve that problem
 
Something like Java's abstract?
 
@AlexA. TIL sometimes strawpoll.me lets me vote three or four times if I'm on mobile.
 
3:30 AM
kind of
Kotlin has a great language design: only things classes define as "overridable"can be overridden
because subclassing is an API, and you need to be able to define that API
 
@DrGreenEggsandHamDJ Haha that's been on my "I'll do this eventually" list for a while. I'd say don't worry about it for now but thanks for reminding me. I'll confer with the mods because we might do a historical lock for some of them.
 
@Geobits ;)
 
@NathanMerrill Kotlin looks really neat. I haven't played with it much yet though.
@Geobits ಠ_ಠ
spits out whiskey that's full of lies and deceit
 
so, taking inspiration, I'd say that if a class has an overridable attribute, and a subclass doesn't override it (or mark it as use default or whatever), then its an error (warning?)
 
I voted once, so there's two at least
 
3:33 AM
@NathanMerrill That could work.
 
@AlexA. "Don't worry about it" I want to worry about it. I want to moderate without the diamond.
 
@QPaysTaxes hence the "use default" part
 
That's why I'm bringing it up.
 
@DrGreenEggsandHamDJ The thing with that is we're going to take pretty individualized actions for most of them. Some might get retagged, some closed, some locked, some deleted.
So it's hard to say "yes let's take action asap" when the right action depends on the post
 
@QPaysTaxes nah, you would be able to define the defaults either in the super class or the sub class
 
3:35 AM
Well, some of them I can help with.
 
@AlexA. Poor whiskey... It's not its fault. Blame me if you want, but don't abuse the whiskey.
 
should I use def, function, func, or something else to define a function?
 
funk
 
@QPaysTaxes by default it would. I would likely use "optional overrideable" or something else to not require the subclass to override
 
Like, this is a pretty obvious "delete"
 
3:36 AM
@Downgoat I think either def or func for brevity.
 
@Downgoat I'd stick with "shred", keeping with the language theme
 
@DrGreenEggsandHamDJ It's already gone. ;)
 
Wow, that was fast.
 
@El'endiaStarman s/brevity/code golf/
 
I just voted on two more.
 
3:37 AM
@Downgoat that said, I thought you were using the () => syntax?
 
@NathanMerrill I'm doing to add both
=>/-> are labdas
 
@Downgoat Haha. Seriously though, it takes a bit longer to type function compared to func and there's more chance of misspelling it.
 
@DrGreenEggsandHamDJ One is both are gone
 
@QPaysTaxes word choice is coming later
because naming is super important
I want to spend much more than 3 seconds
"optional overridding" sounds great to me
 
Wow, do mobiles switch IP addresses that fast? I've voted on this strawpoll five times now, just by waiting a minute or few between votes. Is something else going on?
 
3:39 AM
I'm not even defining the syntax yet
@Geobits are you on wifi or network?
 
Wifi
 
@Geobits So that's why "drink alone" has 6 votes?
 
I'd guess something else then
 
I'd think it wouldn't work.
@AlexA. >_>
 
Bad Geobits, trying to get AlexA. drunk!
 
3:40 AM
Oh no, this is bad. Alex is drunkenly deleting questions.
 
What are you trying to weasel out of him?
 
He's seen me drunk in chat, so it's only fair.
 
Did he get you drunk? o.O
 
( ͡° ͜ʖ ͡°)
 
I'd rather not talk about it...
:P
 
3:42 AM
( ͡° ͜ʖ ͡°)
 
Windows Bash can now also run linux GUI applications, something that wasn't intended by MS, but is awesome anyway: pcworld.com/article/3055403/windows/…
 
@El'endiaStarman drinking != drunk
 
Wow, italic lenny!
 
Italic Lenny looks great in Android chrome.
 
3:43 AM
As in OS X Safari
 
@mınxomaτ Wow, that's cool.
 
ok, another question: On a hashmap, if I put an object where one already exists, should it throw an error?
and introduce another method replace
 
@NathanMerrill yes
 
@mınxomaτ the comments lol
 
or, even better, have put have an option allowReplacements=False
 
3:45 AM
@mınxomaτ there was a pizza ad on that page and now I'm hungry ;_;
 
@QPaysTaxes on what?
oh, I'm definitely doing options
options are actually separate from parameters
and you can set an option on a class, and each time you call a method, it'll use the option you set
 
options just sound like constructor parameters
 
kind of
@QPaysTaxes on methods, the syntax will likely be similar to func(param, param; option1=4, option2=3)
oh yeah, the other thing is that options have to be named
 
@NathanMerrill That's how Python does it. Except the ; is just a ,.
 
3:51 AM
python is a bit different, because it doesn't make a distinction between named parameters and options
but yes, it is definitely similar to python
options actually have to have defaults
anybody else have any cool bugs?
or not cool
I'm not really picky
and the cause?
 
async issues are a pain in javascript
 
@NathanMerrill oh, yes.
 
@Quill I'm actually a fan of how Javascript deals with async.
 
I was trying to debug some PHP code... turns out I spelt verify wrong
I spent TWO weeks debugging that...
 
@NathanMerrill depends what you're doing
 
3:57 AM
;_;
 
@Downgoat IDE?
 
no, just text editor
there is no godod PHP IDE for mac (sorry for the bad enlgish)
 
> godod
 
@Downgoat PHPStorm
@Quill I like the deferred objects, but it can easily lead to massive pyramids of doom
 
ES7 has the await keyword now anyway
 
3:59 AM
@Quill it's ES7
 
Err, ES7 == ES2015?
 
Err, maybe?
 
@El'endiaStarman in 6 to 8 weeks, actually
 
What's gonna happen in 2000 years when ES2015 comes out? :P
 
@QPaysTaxes yes and no, they're still implementing the features in the major browsers
 
4:02 AM
@QPaysTaxes yes
 
@El'endiaStarman lets hope we aren't using Javascript in 2000 years. So much wasted time working with JS
 
@El'endiaStarman it'll be like windows 9, it won't
 
@NathanMerrill Well, y'know, Fortran is still around...
 
There's no reason Fortran shouldn't be used anymore. There is no other scientific language with better compilers available.
 
Hello @Mego!
 
Anonymous
4:06 AM
@QPaysTaxes In its defense, it should've come out in 200x. The committee spent way too much time arguing over concepts
 
@mınxomaτ And modern Fortran is actually a pretty nice language.
 
Anonymous
Tbh I really wish concepts had made it into C++11
 
@Mego they look like interfaces
(just looked them up)
 
@CᴏɴᴏʀO'Bʀɪᴇɴ ಠ_ಠ
 
Anonymous
@NathanMerrill Yeah, formal interfaces are really the only thing C++ is lacking in the OOP side of things
 
4:17 AM
yeah. I actually really like how C++ automatically identifies interfaces
 
Anonymous
You can just do a pure virtual class with all public methods and no instance vars and use multiple inheritance, but that's messy
 
but, that said, enforcing certain requirements would help for future compatability
 
Anonymous
Really anything involving multiple inheritance is messy
 
Anonymous
I would love to have better reflection abilities in C++
 
Anonymous
Like being able to inspect the members of a class
 
4:19 AM
really? I really am not a fan of reflection in static languages
 
Anonymous
But as it is, with name mangling, you have to go through the ABI for that
 
Anonymous
And GCC is the only compiler I know of that has a decent ABI library
 
Anonymous
Yeah GCC does, but I don't think MSVC does
 
Anonymous
Clang might
 
Anonymous
Ideally that would be in the stdlib, with compilers providing their own implementations
 
Anonymous
4:22 AM
Ooh Nathan wanted to hear about bugs
 
Anonymous
I can tell you plenty about bugs
 
awesome!
 
Anonymous
Anything in specific?
 
I'm looking for the reason they existed
 
Yeah, Nathan is trying to design a language where you don't write bugs (or, at least, you find them immediately).
 
Anonymous
4:24 AM
Well one in particular bites me in the ass all the time
 
@El'endiaStarman that's the ideal. I don't actually think its possible
 
@NathanMerrill Well yeah, but you're trying to get as close as possible.
 
yeah :)
 
Anonymous
In Python, there's this handy little feature with list/tuple indexing, where a[-i] refers to the ith item from the end, starting with -1. I constantly forget that it doesn't keep looping, so [1, 2, 3][-4] doesn't work.
 
heh
well, if I do support negative indexing, then I'll definitely mod the number
 
Anonymous
4:27 AM
Though I'm not sure if implementing it as def __getitem___(self, index): return self._list[index % len(self) would be better
 
Anonymous
Would you also do that with positive indexing? As in, [1, 2, 3][4] == 2?
 
There's a problem with modding the index: you never get an index out of bounds error.
 
oooh
yeah
 
And you can't really do it for negative but not positive.
 
actually, I don't think I'll support negative indexes
 
Anonymous
4:28 AM
Basically you'd be turning all lists into circular lists
 
Anonymous
But maintaining the idea of a head and a tail to be able to prepend and append
 
Anonymous
Which I'm not sure would be better
 
I'll definitely include a way to access elements from the end
but I don't think that negative indexing is the way to do it
 
I think the way that Python does it is best. It is super convenient to be able to get the last element easily.
 
Anonymous
If you don't support negative indices, then you get the fun bug where you don't realize the index you're passing in is negative
 
4:30 AM
@El'endiaStarman right but if you have a.last(), or even a.last(2), then things are easy again
sure
I'll come up with a better name
 
Anonymous
Maybe item and ritem?
 
@Mego This would be good. Forward indexing and reverse indexing should be analogous.
(I.e., don't do L[2] and L.last(2) instead of L[-2].)
 
Anonymous
@AlexA. Why did you give "shitpost in TNB" as an option? You imply that you ever don't shitpost :P
 
>_>
 
Anonymous
If I ever made a mainstream language, I'd go for "don't confuse the newbies" as my primary design consideration
 
Anonymous
4:32 AM
Along with The Zen of Python
 
Anonymous
Meanwhile in Seriously I'm basically breaking every rule of The Zen of Python
 
any other bugs Mego?
 
@Mego That's why I want to do \repeat and \loop instead of \for and \while, respectively, but I just know that if I do that, there will be lots of people going to Stack Overflow asking "How do I do a for loop in Pytek?"...
 
Anonymous
@El'endiaStarman I like how you're optimistic and think people will use it enough to ask about it on SO :P
 
@El'endiaStarman I reccomend using for and while as they are more golfy and people who aleady know programming will understand it faster
 
4:34 AM
@Mego I think I remember that Guido designed python to be the language he'd want to use
 
Anonymous
Yeah, and that's a good way to look at it
 
Anonymous
But I'd also want to not confuse new users, because that's the surest way for people to not use your language (see: Seriously)
 
Anonymous
I'm working on a major overhaul for v2 so that I won't be the only one who understands how it works
 
oooh, integer division.
 
I love how GTK apologizes when it crashes.
 
Anonymous
4:37 AM
Oh that reminds me of the exorbitant amount of bugs that I've had to deal with as a result of IEEE 754
 
@Mego JavaScript uses IEEE 754 too!
 
I'm getting my wisdom teeth pulled in 9 hours. -_-
 
@DrGreenEggsandHamDJ >_>
 
what should 1/2 do?
 
Don't let me post any delirious challenges.
 
Anonymous
4:39 AM
The best way to make math functions newbie-friendly is to either use symbolic math (CAS) or arbitrary-precision arithmetic and rationals
 
Anonymous
@NathanMerrill 0.5
 
@NathanMerrill Is it a strictly typed language?
 
@Downgoat yes
but I can define the operator Int/Int to be Double
 
are there seperate integer and float types or are they together in a number class?
@NathanMerrill I don't think operators should output a different type
what if I were to do:
 
@NathanMerrill Yeah, Paul Graham has said that good programming languages are often those that are designed by their creators for themselves, whereas those languages designed for others are usually poor.
 
4:40 AM
integer and float are definitely different
 
int a = 1
a /= 2
print a
would a become a double? throw an error?
 
definitely not become a double
 
so the output should be 0
 
either syntax error or int
yeah
 
Anonymous
90% of people on the planet will expect 1/2 == 0.5
 
4:42 AM
the problem with making Int/Int = Double is that it makes 4/2 -> 2.0
 
@NathanMerrill I highly do not recommend erroring because if I were to divide 50 by a user input, would I have to check if the user input is a factor of 50 before dividing?
 
@Downgoat it would be a syntax error if int/int-> double
 
Anonymous
I think his idea would be something like: 1/2 == 0.5, but 0.5 gets rounded down to 0 when stored in an int. So, double b = a/2; would work and make b == 0.5, and int c = a/2; would also work, but c == 0.
 
because you can't assign a double into an int
 
$ python
>>> 1/2
0
@Mego ^
 
4:44 AM
Silly python.
 
@Mego I'm never going to autocast
 
Anonymous
implicit narrowing is bad
 
Anonymous
@Downgoat That's Python 2, which has many, many problems
 
there would have to be an explicit typecast to integer to get that to work
 
Anonymous
$ python3
Python 3.4.3 (default, May  5 2015, 17:04:32)
[GCC 4.9.2] on cygwin
Type "help", "copyright", "credits" or "license" for more information.
>>> 1/2
0.5
 
4:45 AM
yeah, integer division is the way to go
 
I'm never going to use python 3 because I'm too lazy to move my finger hoof all the way to shift and then to 3
 
Anonymous
Why do you need shift?
 
@DrGreenEggsandHamDJ Thanks for your help with some of the underhanded stuff. I've closed a couple, locked a couple, deleted a couple. Waiting for input from the other mods on a lot of them.
 
Anonymous
Also 3 should be the default everywhere but code golf :P
 
that said, I'm going to bed :)
thanks for all of your input
 
4:46 AM
Pretty much all IDEs, even IDLE, have a way to comment/uncomment lines automatically.
 
@Mego because I have my numbers swapped. pressing the "3" key is #
 
Anonymous
The only use case where 3 doesn't always perform better than 2 is code golf, and even then 3 sometimes does better
 
Anonymous
@Downgoat I don't think I want to know why
 
Anonymous
@El'endiaStarman IDLE annoys me because shift+tab isn't unindent
 
4:48 AM
@El'endiaStarman I dont't thinnk vim does
 
@Mego Haha, yes! I've gotten into the habit of using Ctrl+[ and Ctrl+] though.
@Downgoat @Doorknob probably knows a way.
 
@Mego IDLE's keyboard shortcuts are infuriating.
@Downgoat Yes it does
 
@Doorknob how indent line in vim?
 
gcc
 
nevermind. 4/3 can lead to unexpected bugs
 
4:49 AM
@Downgoat > indent < unindent
 
Anonymous
@AlexA. IDLE is infuriating in general
 
but if you are forced to cast it to an int, then you know what works
 
Anonymous
@El'endiaStarman Do those actually work? Why the hell isn't IDLE using the most widely-used shortcuts for indent and unindent?
 
@Mego They really didn't put as much work into IDLE as they did into the language.
Especially the interactive debugger. Ugh.
 
When I'm Windows I still prefer writing Python in IDLE to Notepad++ though.
 
Anonymous
4:52 AM
People are silly. This answer has 2 upvotes despite only producing valid output for 3 of the test cases: codegolf.stackexchange.com/a/77755/45941
 
Anonymous
Notepad++ or Atom, depending on how big the project is
 
@AlexA. Is it because of the super-convenient F5 shortcut?
 
@El'endiaStarman That, and autocompletion.
 
Anonymous
@El'endiaStarman No, because Notepad++ uses F5 for run also
 
Anonymous
@QPaysTaxes chew
 
4:53 AM
For some reason, "masticate" came to mind. :P
 
@QPaysTaxes Nosh
 
Anonymous
Welcomee
 
@Mego I don't think I've ever actually used Notepad's F5.
 
Anonymous
@QPaysTaxes sleep(1000*60*60*8);
 
@AlexA. Oh yeah, the tab completion is nice.
 
Anonymous
4:53 AM
@El'endiaStarman Me neither, but it's still there
 
Anonymous
@QPaysTaxes Bitte
 
Anonymous
I don't think I've actually ever used tab completion in an IDE. I type very fast and concentrate a lot while typing, so by the time I mentally acknowledge that the tab-completion dialog has popped up, I'm already done with the token.
 
I normally use tab completion for longish variable/function names.
 
Anonymous
I try to be succinct with my var/fn names, so that's never an issue
 
Well, me too, but sometimes...
 
Anonymous
4:59 AM
matrix_determinant is fine and all, but det is a very common shorthand
 
Anonymous
And if a function/variable name gets really long, that's probably a sign that I need to do some refactoring
 

« first day (1903 days earlier)      last day (2945 days later) »