|
rjmccall posted:pfft, i’m a c++ programmer same. Although these days with all the fun C++17 stuff and template metaprogramming it's starting to look more like rust every day.
|
#
?
Nov 3, 2020 22:36
|
|
|
|
| # ? Jun 12, 2024 12:28 |
|
|
haskell! elixir! someone shouts from the back
|
|
#
?
Nov 3, 2020 23:40
|
|
|
lol at Awful.app makin me triple post
eschaton fucked around with this message at 00:58 on Nov 4, 2020 |
|
#
?
Nov 4, 2020 00:48
|
|
|
more Awful excitement eschaton fucked around with this message at 00:59 on Nov 4, 2020 |
|
#
?
Nov 4, 2020 00:49
|
|
|
animist posted:what is dart's raison d'etre anyway keep Gilad Bracha & friends busy and not employed by competitors Strongtalk was great and Newspeak was genuinely innovative, Swift takes some useful cues from it
|
|
#
?
Nov 4, 2020 00:55
|
|
|
oldthinkers unbellyfeel swift
|
|
#
?
Nov 4, 2020 01:01
|
|
|
Talorat posted:same. Although these days with all the fun C++17 stuff and template metaprogramming it's starting to look more like rust every day. c++ is turning into rust, poorly, with a lovely turing complete type system and a subset that’s pre-compiled and you still don’t get memory safety
|
|
#
?
Nov 4, 2020 01:11
|
|
|
Malcolm XML posted:c++ is turning into rust, poorly, with a lovely turing complete type system and a subset that’s pre-compiled and you still don’t get memory safety have pity on the c++ devs. they know all that, but they’ll never get to use rust for real. let them be excited in peace
|
|
#
?
Nov 4, 2020 02:00
|
|
|
elixir lol
|
|
#
?
Nov 4, 2020 02:49
|
|
|
Soricidus posted:oldthinkers unbellyfeel swift
|
|
#
?
Nov 4, 2020 18:35
|
|
|
Malcolm XML posted:to keep the people who wrote v8 employed at Google it didn't even work, lars bak quit to work for an internet of poo poo startup
|
|
#
?
Nov 4, 2020 20:37
|
|
|
b0lt posted:it didn't even work, lars bak quit to work for an internet of poo poo startup it worked for long enough for Google's purposes, I think
|
|
#
?
Nov 4, 2020 20:56
|
|
|
Malcolm XML posted:c++ is turning into rust, poorly, with a lovely turing complete type system and a subset that’s pre-compiled and you still don’t get memory safety to be fair, rust's type system is also turing complete also, extremely vague question: I remember seeing an article about how a small number of compiler optimizations are responsible for like 80% of the performance improvement from a compiler. But now I can't find the article again. Does anybody know of resources that show the relative performance improvements from different compiler passes? Google is failing me.
|
|
#
?
Nov 5, 2020 19:51
|
|
|
it really depends on your code, and your language, and a lot of optimizations are substantially "enabling" optimizations that unblock other optimizations like a ton of c++ code is written in a way which will simultaneously pin values into memory and block almost all memory optimization and do a million tiny calls but after inlining the call overhead goes away and you can just do normal data-flow optimization again. if you tried to comparatively rank optimizations by disabling various optimizations one at a time i'm sure inlining would come in tops but most of its value is not eliminating call oveheads
|
|
#
?
Nov 5, 2020 20:26
|
|
|
I mean compilers are basically a solved problem so why don’t you just let me write whatever I want and then you optimize it for me magically 
|
|
#
?
Nov 6, 2020 05:26
|
|
|
optimization is just inlining wisely
|
|
#
?
Nov 6, 2020 05:30
|
|
|
i could write already optimal code but i choose to allow the compiler to help me
|
|
#
?
Nov 6, 2020 05:58
|
|
|
DELETE CASCADE posted:I mean compilers are basically a solved problem so why don’t you just let me write whatever I want and then you optimize it for me magically big Itanium energy there also since we're talking about optimisations, unspecified argument evaluation order in C++ is one of those that makes some sense in retrospect but is probably going to bite you in the rear end professionally at some point
|
|
#
?
Nov 6, 2020 06:34
|
|
|
we need to abandon the prescriptivist approach to language standardisation. if in practice c++ code widely assumes a certain order of evaluation, specify that poo poo.
|
|
#
?
Nov 6, 2020 11:57
|
|
|
if im gonna waste my weekend learning a functional lang which one should it be. does it even matter at that point?
|
|
#
?
Nov 6, 2020 17:50
|
|
|
Bored Online posted:if im gonna waste my weekend learning a functional lang which one should it be. does it even matter at that point? learn lisp, remember to be extra annoying about it
|
|
#
?
Nov 6, 2020 17:56
|
|
|
Bored Online posted:if im gonna waste my weekend learning a functional lang which one should it be. does it even matter at that point? haskell
|
|
#
?
Nov 6, 2020 17:59
|
|
|
Bored Online posted:if im gonna waste my weekend learning a functional lang which one should it be. does it even matter at that point? java8+
|
|
#
?
Nov 6, 2020 18:23
|
|
|
Bored Online posted:if im gonna waste my weekend learning a functional lang which one should it be. does it even matter at that point? no languages are functional op
|
|
#
?
Nov 6, 2020 18:27
|
|
|
if you're eager to learn a dysfunctional language however...
|
|
#
?
Nov 6, 2020 18:30
|
|
|
Bored Online posted:if im gonna waste my weekend learning a functional lang which one should it be. does it even matter at that point? What are you trying to gain out of learning? If youre not into the type theory and poo poo Clojure is neat, it was the first language that made immutable data structures really click for me. If you are interested in type stuff then haskells obviously the popular pick. But if you wanna go even further off the deep end Idris is neat too.
|
|
#
?
Nov 6, 2020 19:08
|
|
|
Bored Online posted:if im gonna waste my weekend learning a functional lang which one should it be. does it even matter at that point? Standard ML
|
|
#
?
Nov 6, 2020 21:10
|
|
|
It's probably been over 6 months since I mentioned it here so I'm gonna advocate for Erlang
|
|
#
?
Nov 6, 2020 21:18
|
|
|
sml is a very good choice for learning typed functional programming by itself without a whole lot of orthogonal complexity, like monads and type classes in haskell or the sortof scattershot multi-paradigm-ness of ocaml. and the structure/functor system is theoretically neat even if it's somewhat frustrating as an actual language tool
|
|
#
?
Nov 6, 2020 21:18
|
|
|
Tanners posted:What are you trying to gain out of learning? not entirely sure. i did some elixir months back and had an eyes widening experience but it was the first time i looked at anything functional. id like to see what functional programming is like in the practical sense beyond like number and list operations
|
|
#
?
Nov 6, 2020 22:22
|
|
|
i think sml and racket are the places to look at here. sml as the easy intro to cooler type systems, racket as the look at lisps, to take a look at the cool stuff people get up to with dsl's in a context where they are very natural (e.g. stuff like miniKanren)
|
|
#
?
Nov 6, 2020 22:41
|
|
|
Bored Online posted:practical sense
|
|
#
?
Nov 6, 2020 22:51
|
|
|
elm is the best learning language for the same reasons that make it a bitch of a 9-5 language: - it will force you to write code in exactly one design: purely functional mvu. there is no issue with choosing a framework, because there are no frameworks. anything but the default architecture will not compile. makes python look like perl in that regard - it will absolutely never let you take any shortcut. every time your data models are poorly designed, the compiler will punish you by forcing you to handle corner cases you don't care about, until you learn to write tighter models with fewer possible states - it will never provide you with any form of hidden compiler magic, including reflection. you can codegen some trivial stuff like json decoding, but every line of elm code you write you will understand exactly what it does - it has a bus factor of 1 which is 1 higher than a learning language strictly needs
|
|
#
?
Nov 6, 2020 23:08
|
|
|
rjmccall posted:it really depends on your code, and your language, and a lot of optimizations are substantially "enabling" optimizations that unblock other optimizations ty, makes sense. trying to decide whether I should bring llvm into my toy project or just generate lovely code by hand. i'll prolly do it by hand, seems more fun animist fucked around with this message at 23:48 on Nov 6, 2020 |
|
#
?
Nov 6, 2020 23:44
|
|
|
sml is my favorite programming language that i do not expect to ever use again
|
|
#
?
Nov 7, 2020 01:40
|
|
|
maybe I should take a look at sml sometime. I like ocaml a lot but the fact that it does not have a proper multi threading story a decade after multicore became mainstream is just pathetic
|
|
#
?
Nov 7, 2020 02:07
|
|
|
if you've used ocaml then there is probably nothing much to be learned from sml. ocaml has mostly the same concepts but with more practical concessions. since neither actually are practical choices for writing real software i prefer the simpler and more pure one.
|
|
#
?
Nov 7, 2020 02:17
|
|
|
Bored Online posted:if im gonna waste my weekend learning a functional lang which one should it be. does it even matter at that point? nothing matters
|
|
#
?
Nov 7, 2020 02:54
|
|
|
Bloody posted:haskell why tho
|
|
#
?
Nov 7, 2020 03:00
|
|
|
|
| # ? Jun 12, 2024 12:28 |
|
|
Carthag Tuek posted:nothing matters 
|
|
#
?
Nov 7, 2020 03:00
|
|