Return Styles: Pseud0ch, Terminal, Valhalla, NES, Geocities, Blue Moon.

Pages: 1-

If Haskal is that good then...

Name: Anonymous 2016-10-01 18:45

...then why does GHC requires Perl 5? The answer is here1:

I ​noticed that this adds 30.000 lines of C and 1700 lines of Haskell. Not sure if that's any better than just bundling Perl on Windows + a 300 line Perl script.
_________________

[1] - https://ghc.haskell.org/trac/ghc/ticket/9832#comment:12

Name: Anonymous 2016-10-01 19:10

Man, this is hilarious. It's a regexp-based interpreter used at runtime. Written in Perl! In the middle of the Haskell runtime! And they dare claim any sort of type-safety and purity, the liars!

Name: Anonymous 2016-10-01 19:14

sage

Name: Anonymous 2016-10-01 19:16

Here's the script:
https://github.com/ghc/ghc/blob/master/driver/split/ghc-split.lprl

Name: Anonymous 2016-10-01 19:33

And what now? 300 lines of absolutely unreadable shit.

Name: Anonymous 2016-10-01 19:40

Who cares about Windows users?

Name: Anonymous 2016-10-01 20:48

>>5
If you cannot understand that piece of code you better start looking for a new hobby.

Name: Anonymous 2016-10-01 20:52

>>7
Are you telling us you can understand that code? That's impossible. It's Perl, the write-only language.

Name: Anonymous 2016-10-01 21:10

>>8
Any person who is smart enough to understand non-FIOC languages should be able to understand Perl code just fine. But here is a question for you: how much time a given person would need to understand what's behind IO operations in Haskell? I can't give you an exact estimation but I can tell you one thing, a person can memorize the complete perldoc (including the regex engine and all core modules) before that happens.

Name: Anonymous 2016-10-01 21:19

>>8
If you don't know regex syntax then I guess you have zero problems.

Name: Anonymous 2016-10-01 21:26

>>9
what's behind IO operations in Haskell
System calls, that's what. There is nothing arcane about IO operations in Haskell. A putStrLn acts pretty much the same as an printf. You're just making a fuss out of the fact that IO has been made into a magical type. Well guess what, it's not very magical either:

newtype IO a = IO (State# RealWorld -> (# State# RealWorld, a #))

So at the type system level, an IO action is just a function from a state of RealWorld to another state of RealWorld tupled with some result value. IO is just changing the Universe one small step at a time.

Name: Anonymous 2016-10-01 23:04

Haskell needs to fucking die.

Name: Anonymous 2016-10-02 17:53

>>12
No, Haskell is changing the Real World and will change you too.

Name: Anonymous 2016-10-03 0:28

Not every programming language has to be good at everything.

Name: Anonymous 2016-10-03 1:14

>>14
Why are you backpedaling now, huskelnigger?

Name: Anonymous 2016-10-03 1:27

Fuck Haskell
Clojure forever
Long Live Lisp

Name: Anonymous 2016-10-03 8:23

>>16
Clojure is not a Lisp.

Name: Anonymous 2016-10-03 15:28

>>13
you mean like in making the real world pure so then haskell would not have to deal with side effects and other states annoyances?

Name: Anonymous 2016-10-03 17:19

java + xml is GOD
haskal is JOKE

Name: Anonymous 2016-10-03 18:28

>>18
IO is already pure. It is a pure function from one state of the Universe to another. For example, putStrLn is a function from a world without a line output to stdout to a world where this line has been output to stdout.

Name: Anonymous 2016-10-03 18:34

>>20
But pure functions cannot mutate their arguments. So putStrLn only returns a modified copy of the world. There has to be an implicit extra step where you discard the existing world and replace it with the modified copy returned by putStrLn.

Name: Anonymous 2016-10-03 20:22

>>20
heh, it took at least 20 years to develop monadic programming so that languages like Haskal can be ``pure'' and useful, and then one or two decades more to create FRP so I can create a pure version of Pong. The amount of theory behind all this shit is astonishing.

Name: Anonymous 2016-10-03 22:25

>>1

So 300 lines of Perl worth 1700 lines of Haskell and 30.000 lines of C? And I thought Perl is a dead language...

Name: Anonymous 2016-10-04 0:08

>>23
have you ever tried to do any (non-trivial) text processing with Haskell base? it's a serious pain in the butthole; if you are doing some serious job you could just use some pcre binding... but even then, as you can see, the GHC team preferred to stop worrying and use good ol' Perl.

Name: Anonymous 2016-10-04 7:24

>>24
That's sensible given that Perl's strength is in processing text data.

Name: Anonymous 2016-10-04 12:01

>>22
These grants applications ain't gonna write themselves.

Name: Anonymous 2016-10-04 12:04

What programming language is this?

Name: Anonymous 2016-10-04 14:56

>>22
heh, it took at least 20 years to develop monadic programming
That's plain old denotational semantics for procedures.
http://www.di.unipi.it/~andrea/Didattica/PLP-15/DOCS/Tennent-DenotationalSemantics.pdf

Those Haskell guys reinvented the wheel after 20 years of saying wheels are bad.

Name: Anonymous 2016-10-04 15:02

At least Haskell isn't as bad as LISP

Name: Anonymous 2016-10-04 19:55

Whoever has overwritten a Universe with Haskell, please restore it to the old state. I understand that you need to do IO and perhaps even call unsafePerformIO, but I was getting it on with a girl and then she disappeared after your latest update of RealWorld.

Name: Anonymous 2016-10-04 20:41

>>30
after your latest update of RealWorld.
It was not updated. It was garbagecollected and the new world was created.

Name: Anonymous 2016-10-05 8:43

>>31
So GHC is... God?

Name: Anonymous 2016-10-05 12:26

>>32
Godly Haskell Compiler

Name: Anonymous 2016-10-05 15:35

>>30
Do you have any pics of the girl?
Don't change these.
Name: Email:
Entire Thread Thread List
All trademarks and copyrights on this site are owned by their respective parties. All comments and posts are the responsibility of their own posters.
- Tinychan + 2ch Mode + RSS -