December 2018 - Haskell-Cafe

Feedback on plan for checking code requirements in a GHC plugin
by Chris Smith 03 Apr '19

03 Apr '19

Hey everyone, I'm converging on a design for a new feature of CodeWorld, the Haskell-based educational programming environment that I teach with. I'm wondering if anyone has done something like this before. The goal is that when I give an assignment in a class (e.g., "modify this starting code to generalize the repeated pattern using a function"), I want the user to see a checklist of assignment requirements when they run the code. A prototype implementation is here: https://code.world/haskell#PpwpTF1wv3qIvwhohCfvSrQ There are all sorts of possible requirements, such as: "No lines longer than 80 characters", or "there must be a function called foo", or "all top-level definitions must have type signatures", or "your code must define at least 10 top-level variables, and use at least 3 where clauses", or "your definition of var should be equivalent to this one" (see Joachim Breitner's inspection testing work), or even "the function f must satisfy this QuickCheck property". It's been suggested that the requirements language also include the ability to match patterns in the AST, which I think is a good idea. The current prototype uses a pre-compile step that parses the code using haskell-src-exts, and doesn't implement dynamic requirements (runtime-evaluated) at all. My ultimate plan, though, is to send these requirements to GHC via a plugin, then have it evaluate the static ones at compile time, and generate code to check the dynamic ones. Finally, the plugin would add new code to the beginning of main that will invoke a configurable function with the results of the requirement check (hard-coding the static ones, and evaluating dynamic ones on the fly). (In the CodeWorld environment, this function would display the checklist in the web UI, for example.) Has anyone done anything like this before? Any wisdom to share, or ideas to contribute? Thanks, Chris

3 3

role constraints
by Ben Franksen 13 Feb '19

13 Feb '19

I have a question about roles for type variables. I have a number of data types that more or less represent partial bijections. They take two type parameters that represent domain and range. The phantom type parameters are extremely useful: code that shuffles these things around will fail to type check if domains and ranges don't match up exactly. However, there are some situations where I know more about the semantics than the type checker and need to be able to coerce the phantom parameters. No problem, that's what coerce is for... Except, most of the code is polymorphic over the data type, requiring only a few class constraints. And this polymorphism forces the type checker to assume representational roles for the type parameters. Which means I can't use coerce. Instead I have to use unsafeCoerce which I would like to avoid. Is there a way to express, as a constraint, that a certain parameter has a phantom role? What I want to say is: you are only allowed to instantiate the type variable 'a :: * -> * -> *' with data type 'A' if 'x' and 'y' in 'A x y' have both role phantom. And what I get is that I can call coerce to modify the parameters 'x' and 'y'.

2 2

Let's plan BayHac 2019!
by Dan Burton 04 Feb '19

04 Feb '19

It's time to start thinking about the next BayHac! I am assembling a group of organizers for next year's event. Anyone who has been involved in organizing past BayHac events will be automatically accepted. (I have BCCed a number of past organizers; please reach out to any you know in case I missed anyone.) Anyone else who is interested in helping to plan BayHac, please contact me. Depending on how many past organizers rejoin, we will discuss the applicants and accept 1 to 3 new organizers. BayHac is a weekend of learning, sharing, and hacking on a variety of projects for the Haskell community. BayHac usually occurs in San Francisco, California, Mountain View, California, or thereabouts. I'd like to aim for April/May 2019, though that is, of course, subject to change. A rough outline, in rough chronological order, of what the BayHac organizers will be doing: 0. Assemble the organizers. 1. Select date and venue for BayHac 2019. 2. Secure corporate sponsors to cover costs (mainly, food). 3. Invite talk proposals and select speakers. 4. Open online registration and publicize the event. 5. Plan details like catering, registration desk, recording, and make it happen! If any of that sounds like fun to you, then I'd love to hear from you! -- Dan Burton

3 3

HDBC database: error on accent table name
by Damien Mattei 16 Jan '19

16 Jan '19

Hi, i have this error: *** Exception: SqlError {seState = "", seNativeError = 1064, seErrorMsg = "You have an error in your SQL syntax; check the manual that corresponds to your MariaDB server version for the right syntax to use near 'es where Nom = 'A 20'' at line 1"} when doing this : rows_coordonnees <- quickQuery' conn "select * from sidonie.Coordonnées where Nom = 'A 20'" [] it seems tha the tabel name: Coordonnées that contain an accent is causing serious problem to the parser at some point, if i use a table name without accent it works fine. i'm at the point to rename the table which have great impact on all the project build with many other languages (Scheme) that deal correctly the table name with accent. any idea? to make accent works with haskell. Regards, Damien -- Damien.Mattei(a)unice.fr, Damien.Mattei(a)oca.eu, UNS / OCA / CNRS

3 2

MVar considered harmful
by Станислав Черничкин 02 Jan '19

02 Jan '19

Recently I had an interesting discussion on MVars with cats-effect library designers. Cats-effect brings MVar synchronization primitive along with other IO stuff to the Scala programming language. I tried to persuade them to include some Control.Concurrent.MVar’s functions to the library but has failed completely. Moreover, now I think that MVar is a poor choice for basic synchronization primitive. Your may find discussion here https://github.com/typelevel/cats-effect/issues/451 and event try to advocate, tl;dr. Anyway, what is so wrong with MVar? 1. It’s complex. Each MVar has 2 state transitions, each may block. 2. It does not play well in presence of asynchronous exceptions. More specifically, `take` and `put` operations should be balanced (each `take` must be followed by `put`) this force programmer to mask asynchronous exceptions during the MVar acquisition and since `take` function may block, this will delay task cancelation. Well, you may argue what the `takeMVar` function is actually interruptible, but I’m going to show an easier approach which renders interpretability magic unnecessary. What could be the sensible alternative? Guy from the cats-effect suggested me IVar + atomic reference (IORef). This pattern separates concern of blocking (synchronization) from the atomic mutation. So everything can be represented as atomic reference with IVar inside. Just look at this beautiful mutex implementation https://github.com/ovotech/fs2-kafka/blob/master/src/main/scala/fs2/kafka/i… (By ”beautiful” I mean approach itself of course, but not the Scala’s syntax. Scala is one of most ugliest girls after C++ I was forced to date with by my employer for money. Thankfully he didn’t force me to do the same things with her grandma Java). For people who don’t read Scala, the approach is fairly simple. Each thread which want to touch mutex, will create IVar, atomically swap it in the IORef masked (note, that IORef’s operations non-blocking), unmask and wait for previous become available IVar *unmasked*. Then it will either perform it’s operations or fail due to the interruption or exception and trigger newly installed IVar anyway. It just works. Without any «interruptible» magic. So, which benefits can we get? 1. Simpler implementation of basic primitives. Obliviously IORef is fairly simple. IVar is also mind be simpler than MVar, and possibly faster (just “possibly”, I don’t know how it’s implemented, but I guess lesser state transitions implies less logic). 2. Simplified deadlock analysis. Really, we have only IVar with only one state transition and only one potentially blocking operation. 3. Magicless support of interruptions. We don’t need to separate mask/uninterruptibleMask anymore, because all updates are non-blocking, and all waits are unmasked.

10 15

Jan 2019- Membership of the Group
by William Fearon 31 Dec '18

31 Dec '18

1 0

A few questions about unsafe IO
by Qingbo Liu 31 Dec '18

31 Dec '18

Dear Cafe, I am reading an article on HaskellWiki about unsafe IO[1]. It gives the guideline about usage of unsafeDupablePerformIO: "If you need extra speed, and it's acceptable for the action to be performed multiple times, and it's acceptable if this action is canceled halfway through its execution, use unsafeDupablePerformIO.” Inside `ByteStirng` module[2], I noticed that when converting [Char] to ByteString, it uses unsafeDupablePerformIO to allocate space for the ByteString, as the following code shows (important information highlighted): packChars :: [Char] -> ByteString packChars cs = unsafePackLenChars (List.length cs) cs : unsafePackLenBytes :: Int -> [Word8] -> ByteString unsafePackLenBytes len xs0 = unsafeCreate len $ \p -> go p xs0 where go !_ [] = return () go !p (x:xs) = poke p x >> go (p `plusPtr` 1) xs unsafeCreate :: Int -> (Ptr Word8 -> IO ()) -> ByteString unsafeCreate l f = unsafeDupablePerformIO (create l f) -- | A way of creating ByteStrings outside the IO monad. The @Int@ -- argument gives the final size of the ByteString. -- | Create ByteString of size @l@ and use action @f@ to fill it's contents. create :: Int -> (Ptr Word8 -> IO ()) -> IO ByteString create l f = do fp <- mallocByteString l withForeignPtr fp $ \p -> f p return $! PS fp 0 l {-# INLINE create #-} -- | Wrapper of 'mallocForeignPtrBytes' with faster implementation for GHC -- mallocByteString :: Int -> IO (ForeignPtr a) mallocByteString = mallocPlainForeignPtrBytes {-# INLINE mallocByteString #-} The doc of `mallocPlainForeignPtrBytes`, however, explicitly says that no finalizer is added for the allocated memory. So my question is: would not the allocation code in ByteString module cause memory leaks? The doc of `unsafeDupablePerformIO` mentions that it "duplicated IO actions is only run partially, and then interrupted in the middle without an exception being raised”. Thus, it might happen that we have already allocated the memory but then the action is interrupted, without reclaiming the memory. [1] https://wiki.haskell.org/Evaluation_order_and_state_tokens [2] http://hackage.haskell.org/package/bytestring-0.10.8.2/docs/src/Data.ByteSt… Best Regards, Qingbo Liu

2 2

Taking over the udev package
by Leif Warner 30 Dec '18

30 Dec '18

I'm offering to take over the "udev" package on hackage. I have some code which depends on it, as well as a PR to allow it to build on GHC 8.4, among other things: https://github.com/pxqr/udev/pull/5 It looks like the author, Samvel Truzyan, has been inactive on GitHub and Twitter for the past couple of years. Ilya Portnov might also be a good candidate to maintain this package - Ilya has the wacom-daemon package on hackage which depends on udev, and also has previously submitted a PR which has been merged to the udev package. -Leif Warner

1 0

Fwd: IO monad use
by Damien Mattei 29 Dec '18

29 Dec '18

---------- Forwarded message --------- From: Damien Mattei <damien.mattei(a)gmail.com> Date: Sat, Dec 29, 2018 at 9:46 AM Subject: IO monad use To: Damien MATTEI <damien.mattei(a)gmail.com> again an annoying error with my code, i want to apply sort of MAP on list resut of my database IO accessed extracting info with another query, queries works both but i can not MAP or if i can i do not know how to SHOW the result, here is the code: lstNamesBD <- mapM (\(Only name) -> do let nameStr = Text.unpack name bd <- getBD conn nameStr (nameStr,bd)) names it fails to compile with this error: Prelude> :load UpdateSidonie [1 of 1] Compiling Main ( UpdateSidonie.hs, interpreted ) UpdateSidonie.hs:203:33: error: • Couldn't match type ‘(,) String’ with ‘IO’ Expected type: IO (Maybe Float) Actual type: (String, Maybe Float) • In a stmt of a 'do' block: (nameStr, bd) In the expression: do let nameStr = unpack name bd <- getBD conn nameStr (nameStr, bd) In the first argument of ‘mapM’, namely ‘(\ (Only name) -> do let nameStr = ... bd <- getBD conn nameStr (nameStr, bd))’ | 203 | (nameStr,bd)) | ^^^^^^^^^^^^ Failed, no modules loaded. if i code like this show does not know how to display result: let lstNamesBD = Prelude.map (\(Only name) -> do let nameStr = Text.unpack name bd <- getBD conn nameStr res <- (nameStr,bd) res) names putStr "lstNamesBD =" putStrLn $ show lstNamesBD *Main> :load UpdateSidonie [1 of 1] Compiling Main ( UpdateSidonie.hs, interpreted ) UpdateSidonie.hs:193:16: error: • No instance for (Show (IO (Maybe Float))) arising from a use of ‘show’ • In the second argument of ‘($)’, namely ‘show lstNamesBD’ In a stmt of a 'do' block: putStrLn $ show lstNamesBD In the expression: do conn <- connect defaultConnectInfo {connectHost = "moita", connectUser = "mattei", connectPassword = "sidonie2", connectDatabase = "sidonie"} (rows :: [(Text, Double)]) <- query_ conn "SELECT Nom,distance FROM AngularDistance WHERE distance > 0.000278" (names :: [Only Text]) <- query_ conn "SELECT Nom FROM AngularDistance WHERE distance > 0.000278" let resLstNames = Prelude.map fromOnly names .... | 193 | putStrLn $ show lstNamesBD | ^^^^^^^^^^^^^^^ Failed, no modules loaded. help me please

2 5

Re: [Haskell-cafe] About understanding FunBind in haskell-src-exts.
by Magicloud Magiclouds 29 Dec '18

29 Dec '18

Hi, I can see the case of one FunBind contains multiple Matches, like this one. But I cannot see why the "name" (isJust) is in Match, instead of FunBind. Are there cases other than yours that a few functions having different names should be grouped together? PS: FunBind and Match definitions: FunBind l [Match l] Match l (Name l) [Pat l] (Rhs l) (Maybe (Binds l)) On Sat, Dec 29, 2018 at 6:56 PM Alan & Kim Zimmerman <alan.zimm(a)gmail.com> wrote: > > This is for the case when you have a function like > > isJust Nothing = False > isJust (Maybe _) = True > > This is a single FunBind for isJust, but has two Matches > > Alan > > On Sat, 29 Dec 2018 at 12:25, Magicloud Magiclouds <magicloud.magiclouds(a)gmail.com> wrote: >> >> Hi, >> >> Recently I am playing with haskell-src-exts to help understanding >> Haskell syntax more. Then I got this question. >> >> As top level declarations (the type Decl), TypeSig, PatBind are quite >> straight forward, they mean one item in the source. But `FunBind l >> [Match l] >> `, I could not see it maps to "one item" in the source. >> >> Apparently, the list of Match can hold a few non-related functions >> with different names. Then why a FunBind to hold them together? >> >> -- >> 竹密岂妨流水过 >> 山高哪阻野云飞 >> _______________________________________________ >> Haskell-Cafe mailing list >> To (un)subscribe, modify options or view archives go to: >> http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe >> Only members subscribed via the mailman list are allowed to post. -- 竹密岂妨流水过山高哪阻野云飞 And for G+, please use magiclouds#gmail.com.

1 0