Home freenode/#haskell: Logs Calendar

Logs: freenode/#haskell

←Prev  Next→
Page 1 .. 70 71 72 73 74 75 76 77 78 79 80 .. 5022
— 502,152 events total
2020-09-19 17:20:38 <dmwit> do { args <- getArgs; contents <- readFile (head args); putStr contents }
2020-09-19 17:20:53 <dmwit> You might also like The IO Monad for People Who Simply Don't Care.
2020-09-19 17:21:02 <dmwit> http://blog.sigfpe.com/2007/11/io-monad-for-people-who-simply-dont.html
2020-09-19 17:21:54 <lechner> in an earlier experiment, the assignment args <- getArgs seemed to make a difference. is that possible?
2020-09-19 17:22:05 <lechner> instead of using getArgs straight
2020-09-19 17:22:31 <dmwit> Take a look at my link and you will have a framework for understanding why that made a difference.
2020-09-19 17:22:43 → Thra11 joins (~Thra11@5.1.169.217.in-addr.arpa)
2020-09-19 17:22:46 <dmwit> (Yes, it is possible.)
2020-09-19 17:23:27 hackage uniqueness-periods-vector-stats 0.1.1.0 - A very basic descriptive statistics. https://hackage.haskell.org/package/uniqueness-periods-vector-stats-0.1.1.0 (OleksandrZhabenko)
2020-09-19 17:24:50 <sm[m]> https://paste.tomsmeding.com/PZbIiTyi is the simplest fix, as dmwit said
2020-09-19 17:25:56 → wroathe joins (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net)
2020-09-19 17:26:04 <lechner> i am reading and will be back with more shortly. thanks for being so friendly. don't know why people say haskellers are complicated folks :)
2020-09-19 17:26:46 × fendor quits (~fendor@212095005091.public.telering.at) (Ping timeout: 272 seconds)
2020-09-19 17:26:55 × kish` quits (~oracle@unaffiliated/oracle) (Remote host closed the connection)
2020-09-19 17:27:14 → kish joins (~oracle@unaffiliated/oracle)
2020-09-19 17:27:17 × mpereira quits (~mpereira@2a02:810d:f40:d96:b46b:1e98:8653:4550) (Ping timeout: 272 seconds)
2020-09-19 17:27:35 → ukari joins (~ukari@unaffiliated/ukari)
2020-09-19 17:27:43 <sm[m]> I know! We are all simple folk here
2020-09-19 17:28:05 <sm[m]> tomsmeding: should Edit this paste be Clone this paste ? (Also, links between such pastes could be nifty)
2020-09-19 17:28:10 → teardown joins (~user@unaffiliated/mrush)
2020-09-19 17:28:35 <tomsmeding> it should
2020-09-19 17:28:48 <tomsmeding> in fact I have a todo for that
2020-09-19 17:28:52 <lechner> hi, when reading files, is it better to use strict or lazy variants, please?
2020-09-19 17:28:57 → josh joins (~josh@c-67-164-104-206.hsd1.ca.comcast.net)
2020-09-19 17:29:09 <tomsmeding> I like "Clone", too
2020-09-19 17:29:13 <davean> lechner: depends on what you want to accomplish.
2020-09-19 17:29:26 <dmwit> lechner: Lazy IO is to be avoided in real code. For quick experiments while learning, it is fine.
2020-09-19 17:29:28 <merijn> The answer is: use strict or use conduit :p
2020-09-19 17:29:37 <merijn> (or some other streaming library)
2020-09-19 17:29:39 <lechner> thank you
2020-09-19 17:29:54 <lechner> why is lazy I/O a bad idea?
2020-09-19 17:30:02 <merijn> lechner: It's hard to control cleanup
2020-09-19 17:30:19 <lechner> space leaks?
2020-09-19 17:30:22 <davean> lechner: consider if you're reading lazily while writing to it.
2020-09-19 17:30:24 <merijn> lechner: File descriptors (aka handles) are a limited resource and lazy IO makes it hard to control if/when they are freed
2020-09-19 17:30:24 <davean> lechner: what do you get?
2020-09-19 17:30:28 <dmwit> lechner: https://stackoverflow.com/a/6669453/791604 describes why lazy IO is a bad idea
2020-09-19 17:30:45 <dolio> I think lazy I/O is a good idea. :þ
2020-09-19 17:30:49 × wroathe quits (~wroathe@c-68-54-25-135.hsd1.mn.comcast.net) (Ping timeout: 264 seconds)
2020-09-19 17:30:54 → Saten-san joins (~Saten-san@ip-81-11-153-236.dsl.scarlet.be)
2020-09-19 17:31:00 <Uniaika> dolio: I curse you to hell and back
2020-09-19 17:31:18 × walla quits (47b88ff9@pool-71-184-143-249.bstnma.fios.verizon.net) (Remote host closed the connection)
2020-09-19 17:31:31 <davean> lechner: the idea of spaceleaks with laziness is a bit weird, since strict has to use all the space, lazy only might use all the space.
2020-09-19 17:31:32 × ransom quits (~c4264035@2603:300b:1301:3d00:a40f:9b09:1376:26b9) (Ping timeout: 244 seconds)
2020-09-19 17:31:36 <glguy> If your goal is to write a program that opens a file, computes the SHA-512 hash, and prints that; then the lazy IO version will work the same as the strict IO version but probably use less ram
2020-09-19 17:32:01 <lechner> it's the first line in a bigger project
2020-09-19 17:32:02 <merijn> glguy: But then it'd be very little work to use conduit or some such and be sure of that fact ;)
2020-09-19 17:32:19 <merijn> Not to mention that learning 1 or more streaming libraries is worthwhile anyway since they're amazing
2020-09-19 17:32:22 <dmwit> davean: Sure. But lazy encourages a style of programming that would, in a strict semantics, use much more memory than the idiomatic approach from a strict language.
2020-09-19 17:32:34 <merijn> Beat the shit out of nearly anything I've encountered in any other language for such problems
2020-09-19 17:33:14 <dolio> Streaming libraries are exactly the overcomplicated solution for simple intro examples.
2020-09-19 17:33:27 <dmwit> ^
2020-09-19 17:33:32 <merijn> Sure
2020-09-19 17:33:33 <dolio> And are probably an example of why people think Haskell people are complicated.
2020-09-19 17:33:57 <lechner> actually, i think it's because of category theory
2020-09-19 17:34:12 <davean> dmwit: I mean that sounds a lot like the tautology of "if you used more memory, you'd use more memory".
2020-09-19 17:34:26 <geekosaur> you need zero category theory to use Haskell
2020-09-19 17:34:29 <merijn> lechner: Category theory has basically zero relevance to haskell projects (or even Haskell)
2020-09-19 17:34:51 <merijn> lechner: I dunno where that meme got started on the internet, but it's 100% nonsense
2020-09-19 17:35:16 <dolio> Also note that one of the negatives about lazy I/O brought up (what if you read and write to a file concurrently) also applies to streaming libraries, but no one blames the streaming libraries for it for some reason.
2020-09-19 17:35:49 <merijn> lechner: I've been doing Haskell for, like, a decade now. Very productively even, and I still got lost about 2-3 chapters into any book on category theory :p
2020-09-19 17:35:57 <davean> dolio: uh, doesn't happen when I use 'machines' for example, I can can control the reading and writing so I can know what happens exactly.
2020-09-19 17:36:06 <merijn> dolio: Files are shit, let's replace everything with SQLite databases :D
2020-09-19 17:36:10 <davean> dolio: infact I use 'machines' specificly for reading and writing to a file at the same time.
2020-09-19 17:36:27 hackage texmath 0.12.0.3 - Conversion between formats used to represent mathematics. https://hackage.haskell.org/package/texmath-0.12.0.3 (JohnMacFarlane)
2020-09-19 17:36:56 <lechner> is there a kindred spririt between haskell and the declarative nature of SQL?
2020-09-19 17:37:23 <merijn> Kinda, maybe?
2020-09-19 17:37:46 → Sgeo joins (~Sgeo@ool-18b982ad.dyn.optonline.net)
2020-09-19 17:37:47 <davean> dolio: yah, I'm not sure why you think the concurrency would be an issue with a streaming library.
2020-09-19 17:37:48 <merijn> SQL (at least, the non-write statements) is purely functional, after all :)
2020-09-19 17:37:55 → DavidEichmann joins (~david@43.240.198.146.dyn.plus.net)
2020-09-19 17:37:56 <ddellacosta> merijn: re: category theory in Haskell, I think the unfortunate truth is that as soon as people see words like "Functor" they assume that Haskell is a morass of abstract nonsense and stop thinking there
2020-09-19 17:38:01 × DavidEichmann quits (~david@43.240.198.146.dyn.plus.net) (Remote host closed the connection)
2020-09-19 17:38:22 <sm[m]> conduit is awesome but a bit too much for someone writing their first haskell program
2020-09-19 17:38:35 <merijn> Well, that wasn't really specified :p
2020-09-19 17:38:48 <sm[m]> twas, twas
2020-09-19 17:38:52 <merijn> The question was "lazy or strict" :p
2020-09-19 17:38:54 <dolio> davean: Because streaming libraries don't solve having to not be naive about interleaving your reading and writing to a single file, which is what people blame lazy I/O for.
2020-09-19 17:39:13 <merijn> dolio: Nobody claimed streaming libraries solved "all problems ever" so that seems a bit of a strawman
2020-09-19 17:39:15 <davean> dolio: they can dothat.
2020-09-19 17:39:15 <geekosaur> that's only one of the problems with lazy I/O
2020-09-19 17:39:33 <davean> dolio: for example, only executing the writes to already read parts.
2020-09-19 17:39:52 <davean> and delaying the ones to pending reads.
2020-09-19 17:39:58 <merijn> Also, I don't think anyone thinks "simultaneous read and writes" is the problem with lazy IO, considering none of the streaming libraries even talk about that in any motivation
2020-09-19 17:40:27 <dolio> merijn: Then they should stop using it as an example of lazy I/O being a problem.
2020-09-19 17:40:47 <merijn> dolio: I've literally never seen that brought up as a lazy IO problem except just now on IRC
2020-09-19 17:41:01 <dolio> Well I've seen it multiple times in the past on IRC.
2020-09-19 17:41:04 × dansho quits (~dansho@ip68-108-167-185.lv.lv.cox.net) (Quit: Leaving)
2020-09-19 17:42:08 <davean> dolio: to be clear I brought it up about lazy vs. strict io, not streaming. Streaming systems can't solve problems without you deciding what you want.
2020-09-19 17:42:10 <sm[m]> lechner, please ask less controversial questions in future (j/k :)
2020-09-19 17:42:18 <merijn> Yes
2020-09-19 17:42:27 <merijn> Like "how do I get 'a' out of 'IO a'?" ;)
2020-09-19 17:42:49 <davean> merijn: well, you just pattern match on (# RealWorld, ... #) ...
2020-09-19 17:43:21 <merijn> ;)
2020-09-19 17:43:34 <davean> Or use the Comonad instance of IO
2020-09-19 17:44:04 <davean> which is easy to write with unsafePerformIO
2020-09-19 17:44:07 <davean> laws aside
2020-09-19 17:44:17 <tomsmeding> :D
2020-09-19 17:44:23 × MaoZeDong_ quits (~yuri@2a00:1370:8135:91d4:415b:9b05:36f8:565c) (Ping timeout: 272 seconds)
2020-09-19 17:44:31 <monochrom> http://www.vex.net/~trebla/photo/unorganized/IO-String.png
←Prev  Next→
Page 1 .. 70 71 72 73 74 75 76 77 78 79 80 .. 5022

All times are in UTC.