- Haskell-Cafe - Haskell.org

Integrate GHC-API with Cabal
by Félix Baylac 26 Jul '18

26 Jul '18

Dear cafe, I am currently working on a project [1] which aims to index the code published on Stackage in order to provide a "code example" database. Basically, for each library, I want to generate a real-world example corpus for each exported symbol. Before generating this database, I need to retrieve the exported symbol of a package. So far, I have been gathering the dependencies and integrating them in the GHC pkg database using an external stack call, I then parse the cabal file and load the exposed modules using the GHC API and gather the exported symbols*. It works pretty reliably on the simple packages, however, I end up with some missing dynamic flags for some more advanced packages (missing c includes, c libraries, ASM flags, etc.). I then started to parse and load these missing attributes to GHC until I stepped back for a moment and realized I was re-implementing cabal build. I then looked for a way to "hook" my GHC API code in cabal build. This would allow cabal to both handle the dependencies gathering as well as setting up the correct GHC dyn-flags. The only resource I found was [2]. It's really clever, however, looks a bit hacky to me. Is there a better way to perform this kind of "hook"? [1] https://git.alternativebit.fr/NinjaTrappeur/Exhs [2] http://blog.ezyang.com/2017/02/how-to-integrate-ghc-api-programs-with-cabal/ * I am aware of the hoogle index generated by haddock. However, this index is missing when the haddock documentation of a package cannot be generated.

4 4

Re: [Haskell-cafe] Why does this result in a runtime loop instead of a type error or typechecker loop?
by Anthony Clayden 26 Jul '18

26 Jul '18

On Tue, Jul 24, 2018 at 5:52 AM,Ryan Reich wrote: > This seems like the canonical illustration of the dangers of UndecidableInstances. No I don't buy that explanation. The worst UndecidableInstances can do is send the type checker into a loop. For mutually recursive functions to cause looping, you don't need classes at all: foo :: Bool -> () foo x = bar x bar :: () -> Bool bar x = foo x No typechecking will detect that loop. This `HasBool` example is just an obfuscated way to write that loop(?) There's some odd thigs in the OP: Why use OVERLAPPABLE? Neither of those instances need it. Or are there more instances in the imports? (There must at least be the definition for class `Has`.) If so, what are their constraints? And what are their implementations? Perhaps that's causing the looping? Will `runRIO` or `view` use one of those other instances? What does this mean: "I can compile an equivalent program; it loops on execution."? There doesn't need to be a `Bool` in the "environment": the `HasBool a` instance introduces `Bool` in its constraint. I suggest boiling this down to a stand-alone program not needing imports. Then tell the whole story. AntC On Mon, Jul 23, 2018 at 2:10 PM, Theodore Lief Gannon < tanuki at gmail.com <http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe>> wrote: >* λ. :set -XFlexibleContexts -XFlexibleInstances -XMultiParamTypeClasses *>* -XUndecidableInstances *>* λ. import RIO *>* λ. import Data.Has *>* λ. class HasBool a where boolL :: Lens' a Bool *>* λ. instance {-#OVERLAPPABLE#-} HasBool a => Has Bool a where hasLens = *>* boolL *>* λ. instance {-#OVERLAPPABLE#-} Has Bool a => HasBool a where boolL = *>* hasLens *>* λ. runRIO () $ view boolL *>* ^CInterrupted. *>>* The RIO environment () doesn't contain a Bool. I can see how the *>* typechecker might get lost in this, but I can compile an equivalent *>* program; it loops on execution. *>>* _______________________________________________ *>* Haskell-Cafe mailing list *>* To (un)subscribe, modify options or view archives go to: *>* http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe <http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe> *>* Only members subscribed via the mailman list are allowed to post. *>

2 1

Re: [Haskell-cafe] Integrate GHC-API with Cabal
by Daniel Gröber 25 Jul '18

25 Jul '18

Hi, On Wed, Jul 25, 2018 at 08:30:48PM +0200, Félix Baylac wrote: > Thanks a lot for this really helpful message. > > > I think you might be interrested in one of my packages, > > [cabal-helper](https://hackage.haskell.org/package/cabal-helper). It's > > geard more towards editor tooling but I think it should work for your > > use case too. > > Indeed, looks like what I'm looking for. > > > Basically you'd have to do `cabal install --only-dependencies && cabal > > configure && cabal build` for each package you want to inspect and > > then cabal-helpers API boils down to giving you a list of GHC > > flags/module names you need to pass into the GHC API to make it work. > > > > (The `cabal build` step is needed to run preprocessors like hsc2hs and > > compile C dependencies, but once cabal has taken care of that you can > > usually just load the package without problems.) > > I had to make some modifications and jump in some nix hoops to use > cabal-install instead of stack to build the packages, but indeed, it does work > like a charm. Ah, I forgot you said you were using Stack. You can use it too if you prefer. You just have to use `stack path --dist-dir` (run in the project directory) to get the path to the dist-dir after running `stack build` instead of using the hardcoded `$PWD/dist` value. There have been some reports of issues with stack using weird combinations of Cabal-library versions though which so it might need some tweaking. > This is perfect, I just tried your library following this example on the text > package and it seems to do the job. Thanks a lot, fixing this problem without > re-inventing the wheel is a huge relief. > > Do you mind if I write a small article on this integration. No not at all, send me a link when you're done. > It'll probably prevent some other people from using dodgy hacks. Well I'm not sure I would call cabal-helper not a dodgy hack but at least it's a (sort of) well maintained hack :) --Daniel PS: Looks like I messed up the list-reply on my original message so I resent it and I'm CC'ing haskell-cafe on this one. Sorry about that.

1 0

Haskell Opportunity - New York - Blockchain
by Alex Mesropians 25 Jul '18

25 Jul '18

Hey All, All Haskell engineers in the US we have a new and exciting role with a blockchain firm who are rapidly expanding! Check them out - Haskell Role <http://t.sidekickopen13.com/e1t/c/5/f18dQhb0S7lC8dDMPbW2n0x6l2B9nMJW7t5XZs5…> Thanks, Alex

1 0

how to terminate a (Posix) process and all its subprocesses?
by Johannes Waldmann 24 Jul '18

24 Jul '18

Dear Cafe, I am using https://hackage.haskell.org/package/process-1.6.4.0/docs/System-Process.htm… to start an external command, and wait for its completion. Now I would like to time-out this after a while. I can use https://hackage.haskell.org/package/base-4.11.1.0/docs/System-Timeout.html and this works in simple cases - but not in case the external command has spawned child processes. (cleanupProcess sends SIGTERM but only to the process at the top of the tree ?) I guess I need to use https://hackage.haskell.org/package/unix-2.7.2.2/docs/System-Posix-Process.… but that seems rather low-level - and I don't see how I would get the ProcessID of the process started by readProcess, so I'd also have to re-do that. Is there an abstraction/library that would help here? Thanks - J.W.

6 7

Free theorem for `forall z. (A,z) -> (B,z)`?
by Conal Elliott 24 Jul '18

24 Jul '18

Suppose `g :: forall z. (A,z) -> (B,z)`. Is it necessarily the case that `g = first f` for some `f :: A -> B` (where `first f (a,b) = (f a, b)`), perhaps as a free theorem for the type of `g`? Note that `(A,)` and `(B,)` are functors and that `g` is a natural transformation, so `g . fmap h == fmap h . g` for any `h :: u -> v`. Equivalently, `g . second h == second h . g` (where `second h (a,b) = (a, h b)`). Thanks, -- Conal

4 8

Design of an extensible system
by Marc Busqué 24 Jul '18

24 Jul '18

Hi there, I'm currently programming my first non toy project in Haskell. I come from Ruby background, where everything is possible no matter how much entropy the universe gains after you finish your work, so I'm still not sure how to do certain things without this side-effect :) I'll try to be concise. My idea is that the project would consist of several different Haskell packages. The central one, call it Core, would define a certain data type (with info about certain URLs). Other packages would define each one an strategy to do something with that data type (processing the URL info). Each strategy would have its own dependencies, and as a user probably won't need all of the strategies, I would like to keep them in separated packages so she or he has no need to download the unneeded stuff. The problem is that Core should act as the central dashboard, whether through a CLI or even a web frontend. For example, in the CLI case, I would like to be able to install Core package along with package StrategyA and package StrategyB and then do something like the following: ``` bash> core --action=fetch --strategy=A bash> core --action=fetch --strategy=B ``` Two generic ideas come to my mind but I don't know how to translate them to Haskell: - Infer module name from the `--strategy` argument and call it dynamically. However, as far as I know, in Haskell modules are not first class citizens, so no way I can dynamically call them. I have read something to achieve this through Template Haskell or through outdaded [plugins](http://hackage.haskell.org/package/plugins) package, but I'm not sure whether the need is the same and anyway it seems quite unnatural to Haskell (I would not like to go against it, but to learn to do thing in its way). - Define from Core some kind of global list where strategy packages can register a map between CLI argument value and a defined instance of the core data type. But, again, how to do global things in Haskell (and I would prefer not to need external tools like RDBMS, filesystem or whatever)? And, anyway, surely I don't want to do this because concurrency and everything... Any idea? Thanks in advance! Marc Busqué http://waiting-for-dev.github.io/about/

6 9

Expressing disjunctions of constraints
by Olaf Klinke 24 Jul '18

24 Jul '18

Dear cafe, today I wrote a type class with three parameters. class Foo a b c | a b -> c where foo :: a -> b -> c instance Foo A R A where ... instance Foo R A A where ... In addition to the functional dependency I'd like to express that at least one of the parameters a and b is c, that is, Foo is the union of the two classes class Foo1 a b where foo1 :: a -> b -> a class Foo2 a b where foo2 :: a -> b -> b and furthermore the choice between Foo1 and Foo2 determines one parameter to be a fixed type, say R. I understand that logical disjunction poses problems for instance resolution [1] because adding an instance may affect the instance dictionary lookup. This should not happen in my case because in the end this would be a one-parameter type class. The ultimate aim is to use the same symbol for both a function foo :: R -> a -> a and for flip foo. Of course writing flip foo is not that much of a nuisance, I just wonder whether Haskell is expressive enough to do this. Olaf [1] https://stackoverflow.com/questions/10255148/how-can-i-combine-two-type-con…

3 2

animation application
by Dennis Raddle 24 Jul '18

24 Jul '18

I'm working on an application for animating words, images, and shapes. I'm actually doing it in Purescript and it will become part of a math game that runs in the browser, but I'm asking my question on the Haskell list because the issues seem central to a functional/pure paradigm. Let's say that the things being animated are "objects." Objects could include photos, text, shapes, lines, etc. They can have fixed positions, colors, and other attributes, or their attributes can change over time. How do I compute attributes that change over time? In some cases, a function with a single argument, the current time, would suffice. I might want state, though -- perhaps an object's movement is pseudorandom and needs to refer to a seed and past position state. Objects might even need to observe each other's state. For instance, imagine we have two words, W1 and W2, and they move according to some stateful algorithm. Meanwhile we want to draw a line connecting their centers, so we need a line object L that can read the centers of W1 and W2. How do I represent the state of W1, W2, and L? I could define a general Object data type: (here there are no fields mentioned) data Object = W1 | W2 | L I will need state and the ability to customize the drawing and positioning functions, so to flesh out the fields of, say , W1. type St = StateT WholeScene IO data Object = W1 { leftEdge :: Int , rightEdge :: Int , topEdge :: Int , bottomEdge :: Int , computePosition :: W1 -> Time -> St W1 , draw :: W1 -> St () , ... other state ... } Okay, so now I could define W2, which will be different in some ways, perhaps needing different state. But it will look a lot alike. It might have the same edge fields. In fact, all my objects will have similar position, color, etc. attributes. I certainly don't want to reinvent the wheel for everything. I also want objects to be able to read each other's state. Consider a basic positioning task: some object Obj1 keeps its left edge 100 pixels to the right of some other object's right edge. So the Obj1 function for computing position needs access to the other objects. Also consider that all these objects need to be put into a list together, to hold the state of the entire scene. They need to be members of the same class, and have a way to look each other up. One way to do this is have a single Object type with numerous constructors, and define similar functions on every constructor type (say, a function for accessing or setting the left edge). I also thought about having an Object be an existential type, and the fields that are custom to one object will be the existentially defined data. I actually implemented this in Purescript, but it got out of hand, probably due to my own weakness of understanding how existential types work. Any ideas? D

1 1

Why does this result in a runtime loop instead of a type error or typechecker loop?
by Theodore Lief Gannon 24 Jul '18

24 Jul '18

λ. :set -XFlexibleContexts -XFlexibleInstances -XMultiParamTypeClasses -XUndecidableInstances λ. import RIO λ. import Data.Has λ. class HasBool a where boolL :: Lens' a Bool λ. instance {-#OVERLAPPABLE#-} HasBool a => Has Bool a where hasLens = boolL λ. instance {-#OVERLAPPABLE#-} Has Bool a => HasBool a where boolL = hasLens λ. runRIO () $ view boolL ^CInterrupted. The RIO environment () doesn't contain a Bool. I can see how the typechecker might get lost in this, but I can compile an equivalent program; it loops on execution.

3 2