October 2023 - Haskell-Cafe

Best way to implement "scoped exceptions"?
by Tom Ellis 27 May '25

27 May '25

I'd like to implement "scoped exceptions"[1], that is, a combinator of type withScopedException :: ((forall a. e -> IO a) -> IO r) -> IO (Either e r) so I can use exceptions in a convenient way, without having to set up fragile "action at a distance" based on the type of the thing I'm throwing and catching, for example: scopedExceptionExample :: IO (Either String (Either Int Void)) scopedExceptionExample = do withScopedException $ \throw1 -> withScopedException $ \throw2 -> if (1 :: Int) < 0 then throw1 "Hello" else throw2 1234 -- ghci> scopedExceptionExample -- Right (Left 1234) Here's a hacky way of doing it, based on tagging each exception with a unique value, and then filtering when handling: data MyException where MyException :: e -> Data.Unique.Unique -> MyException instance Show MyException where show _ = "<MyException>" instance Exception MyException withScopedException :: ((forall a. e -> IO a) -> IO r) -> IO (Either e r) wiathScopedException f = do fresh <- Data.Unique.newUnique flip tryJust (f (\e -> throwIO (MyException e fresh))) $ \case MyException e tag -> -- unsafeCoerce is very unpleasant if tag == fresh then Just (unsafeCoerce e) else Nothing This is the approach taken by the effectful library[2]. But is there a better way? Can I persuade GHC's RTS to work like this directly? Tom [1] I don't know if these already have a name [2] https://hackage.haskell.org/package/effectful-core-2.2.1.0/docs/src/Effectf…

5 7

How to use the GPU in Apple M2 from Haskell?
by David Banas 15 Nov '23

15 Nov '23

Does anyone have a good reference for explicitly accessing the GPU in the new Apple M2 chip from a user Haskell program?

2 3

Question on Lexing Haskell syntax
by Travis Athougies 02 Nov '23

02 Nov '23

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512 According to the Haskell report [1] (See Note 5), a virtual `}` token is inserted if parsing the next token would cause a parse error and the indentation stack is non-empty. I'm trying to lex and parse Haskell source and this sort of interplay (which requires two-way communication between lexer and parser) makes it very difficult to write a conformant implementation. I can't change the standard (obviously), but I'm wondering if this is actually what GHC (de facto the only Haskell compiler) does, or if it applies some other rule. If so, does anyone know the exact mechanism of its implementation? I've been programming Haskell for more than a decade, and while I have an intuitive understanding of the indentation rules, I would have assumed the source could be lexed without also having a parser. In particular, the note seems to imply that the main purpose of this is to properly lex `let`/`in` bindings. Perhaps there's an alternate equivalent rule? Curious to hear other's thoughts. Travis [1] https://www.haskell.org/onlinereport/haskell2010/haskellch10.html#x17-17800… -----BEGIN PGP SIGNATURE----- iIsEARYKADMWIQRW81c55hOCXuY5k/H+BxaRGQDTjQUCZUGARhUcdHJhdmlzQGF0 aG91Z2llcy5uZXQACgkQ/gcWkRkA041W+wEA/7n5NejGTYu4O6N+Pt7Rn0bBRw6D 5D96idagahXqXioA/18hxYFpY45lWwB7pKCh83xQJu2Bcwkxj1xhhCEZBcoA =FrDy -----END PGP SIGNATURE-----

5 6

Munich Haskell Meeting, 2023-10-31 @ 19:30
by Leah Neukirchen 30 Oct '23

30 Oct '23

Dear all, Next week, our monthly Munich Haskell Meeting will take place again on Tuesday, October 31 at Iveria (Lindwurmstraße 159A) at 19h30. For details see here: http://muenchen.haskell.bayern/dates.html If you plan to join, please add yourself to this nuudel so we can reserve enough seats! It is OK to add yourself to the nuudel anonymously or pseudonymously. https://nuudel.digitalcourage.de/7ST4pxNRVj3rGLdO Everybody is welcome! cu, -- Leah Neukirchen <leah(a)vuxu.org> https://leahneukirchen.org/

2 1

Re: [Haskell-cafe] Should GHC 9.4.7 be native on MacBook Air M2?
by David Banas 28 Oct '23

28 Oct '23

Ah, I needed this line in my stack.yaml file: arch: aarch64 Why is that? Shouldn’t it use the native architecture if it can support it? -db

2 1

Should GHC 9.4.7 be native on MacBook Air M2?
by David Banas 28 Oct '23

28 Oct '23

I’m working on a new MacBook Air M2. I just bumped my current Haskell project to stack LTS-21.17, after catching up on the GHC porting effort news. I did a stack clean / stack build and saw GHC 9.4.7 being installed, as expected, but still to the x86_64-osx directory: % l ~/.stack/programs x86_64-osx % l ~/.stack/programs/x86_64-osx ghc-8.10.3 ghc-8.10.4.tar.bz2 ghc-8.6.3.installed ghc-9.0.2 ghc-9.2.2.tar.bz2 ghc-9.2.7.installed ghc-8.10.3.installed ghc-8.4.3 ghc-8.6.3.tar.bz2 ghc-9.0.2.installed ghc-9.2.5 ghc-9.2.7.tar.bz2 ghc-8.10.3.tar.bz2 ghc-8.4.3.installed ghc-8.6.5 ghc-9.0.2.tar.bz2 ghc-9.2.5.installed ghc-9.4.7 ghc-8.10.4 ghc-8.4.3.tar.bz2 ghc-8.6.5.installed ghc-9.2.2 ghc-9.2.5.tar.bz2 ghc-9.4.7.installed ghc-8.10.4.installed ghc-8.6.3 ghc-8.6.5.tar.bz2 ghc-9.2.2.installed ghc-9.2.7 ghc-9.4.7.tar.bz2 I was expecting a native version; do I misunderstand something?

1 0

Simplest way to move a ... `using` ... line into the IO monad?
by David Banas 27 Oct '23

27 Oct '23

Hi all, What is the simplest way to translate a non-monadic line of Haskell code, such as: res = map foo xs `using` parList rseq For use in the IO monad? This doesn’t work: res <- mapM fooM xs `using` parTraversable rseq complaining: No Traversable instance for IO. Thanks, -db

4 6

Re: [Haskell-cafe] when does ghc(i) expand type synonyms?
by Olaf Klinke 26 Oct '23

26 Oct '23

> ghci> :set +t > > ghci> "foo" > "foo" > it :: String > > ghci> reverse "foo" > "oof" > it :: [Char] I'm not a GHC expert, but my guess is: It must have something to do with unifying the argument types with the type signature of the function. In (<>), the two argument and the result types are identical. Hence it can be safely said that x <> whatever as long as this expression is valid, has the same type as x. In contrast, the result type of (++) is always of the shape [a]. Hence String gets unified with [a] which forces expansion of the type synonym. Consider: class Foo a where {rev :: a -> a; conc :: a -> a -> a} instance Foo [a] where {rev = reverse; conc = (++)} revconc :: Foo a => a -> a; revconc a = conc a (rev a) :t revconc ("foo" :: String) revconc ("foo" :: String) :: String So as long has you hide (++) in a type signature that proves input and output are the same, synonyms don't get expanded. Olaf

1 0

when does ghc(i) expand type synonyms?
by Johannes Waldmann 25 Oct '23

25 Oct '23

Dear cafe, my students were asking: why are types printed differently here? ghci> :set +t ghci> "foo" "foo" it :: String ghci> reverse "foo" "oof" it :: [Char] ... and I add ghci> fst ("foo", 42) "foo" it :: String compiler docs do not really explain the difference https://downloads.haskell.org/ghc/latest/docs/users_guide/ghci.html#ghci-cm… (it says "display the type of .." while technically it is "display some representation of the type of ..."?) I found this explanation of a somewhat related issue https://gitlab.haskell.org/ghc/ghc/-/issues/9183#note_83743 While playing around with this, I noticed: ghci> let x = "foo" in (x <> reverse x) "foooof" it :: String ghci> let x = "foo" in (x ++ reverse x) "foooof" it :: [Char] Well, it's just optics: when I extract the type programmatically, the synonym necessarily does get expanded ghci> import Data.Data ghci> typeOf (let x = "foo" in (x <> reverse x)) [Char] it :: TypeRep - Johannes

2 1

Hackage not building package
by Ivan Perez 23 Oct '23

23 Oct '23

Hi, I recently uploaded a package ( https://hackage.haskell.org/package/dunai-0.12.0) and hackage is not building it. I don't see info anywhere and I cannot re-trigger a build: attempting to do so shows the message "Report not found -Build report does not exist" Is everything ok with hackage? Thanks Ivan

2 1