- ghc-devs - Haskell.org

Why is an irrefutable pattern desugared to a 'cast', while an incomplete pattern is not?
by Christiaan Baaij 12 May '14

12 May '14

Dear list, When I ask for the desugaring of: > module PatError where > > paterror :: Maybe Int -> Int > paterror (Just i) = i I get the following: > PatError.paterror = > \ (ds_dIS :: Data.Maybe.Maybe GHC.Types.Int) -> > break<1>() > case ds_dIS of _ [Occ=Dead] { > __DEFAULT -> > (\ _ [Occ=Dead, OS=OneShot] -> > Control.Exception.Base.patError > @ GHC.Types.Int "PatError.hs:4:1-21|function paterror"#) > GHC.Prim.void#; > Data.Maybe.Just i_aqG -> break<0>(i_aqG) i_aqG > } Where 'Control.Exception.Base.patError' is applied to 'Int'. However, when I compile the almost identical code with an irrefutable pattern: > module PatError2 where > > patError :: Maybe Int -> Int > patError ~(Just i) = i I get the following core: > PatError2.paterror = > \ (ds_dIT :: Data.Maybe.Maybe GHC.Types.Int) -> > break<1>() > let { > i_aqG :: GHC.Types.Int > [LclId, Str=DmdType] > i_aqG = > case ds_dIT of _ [Occ=Dead] { > __DEFAULT -> > (\ _ [Occ=Dead, OS=OneShot] -> > (Control.Exception.Base.irrefutPatError > @ () "PatError2.hs:4:1-22|(Data.Maybe.Just i)"#) > `cast` (UnivCo representational () GHC.Types.Int > :: () ~# GHC.Types.Int)) > GHC.Prim.void#; > Data.Maybe.Just i_aqG -> i_aqG > } } in > break<0>(i_aqG) i_aqG Where 'Control.Exception.Base.patError' is first applied to '()' after which it is casted to 'GHC.Types.Int'. Out of the [note] for the 'UnivCo' coercion: > The UnivCo ("universal coercion") serves two rather separate functions: > - the implementation for unsafeCoerce# > - placeholder for phantom parameters in a TyConAppCo It seems to be used in the 'unsafeCoerce#' role here. My question is: why can't the irrefutable pattern be translated to 'Control.Exception.Base.irrefutPatError' applied to 'Int'? Instead of using this 'cast'? I find the cast-less version much more readable. Greetings, Christiaan

2 1

How to fix T5979?
by Joachim Breitner 10 May '14

10 May '14

Hi, SPJ recently committed: commit 12332f1add3eb9615a62abb36898819b2e67e865 Author: Simon Peyton Jones <simonpj(a)microsoft.com> Date: Thu May 8 13:06:05 2014 +0100 Error message wibble, presumably due to recent changes in transformers diff --git a/testsuite/tests/ghci/scripts/T5979.stderr b/testsuite/tests/ghci/scripts/T5979.stderr index fe6e012..7230011 100644 --- a/testsuite/tests/ghci/scripts/T5979.stderr +++ b/testsuite/tests/ghci/scripts/T5979.stderr @@ -1,4 +1,4 @@ <no location info>: Could not find module ‘Control.Monad.Trans.State’ - It is not a module in the current program, or in any known package. + It is a member of the hidden package ‘transformers-0.3.0.0’. But that breaks when transformers is indeed not installed (as it is the case in a minimal validation run such as on travis): Actual stderr output differs from expected: --- ./ghci/scripts/T5979.stderr 2014-05-10 03:10:07.860423047 +0000 +++ ./ghci/scripts/T5979.run.stderr 2014-05-10 03:32:54.861642002 +0000 @@ -1,4 +1,4 @@ <no location info>: Could not find module ‘Control.Monad.Trans.State’ - It is a member of the hidden package ‘transformers-0.3.0.0’. + It is not a module in the current program, or in any known package. *** unexpected failure for T5979(ghci) For now I added a reqlib('transformers') to the test, hoping that that helps. But why should the error message mention "transformers" at all when the package import is actually explicit to import "nonexisting"? $ cat T5979.script :set -XPackageImports import "nonexisting" Control.Monad.Trans.State import "transformers" Control.Monad.Trans.State import "transformers" Control.Monad.Trans.Reader I am a bit worried that there is a regression here and SPJs commit just hides it. But gotta run now and cannot investigate further. Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • http://www.joachim-breitner.de/ Jabber: nomeata(a)joachim-breitner.de • GPG-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

3 4

Bugsquashing at ZuriHac
by Joachim Breitner 10 May '14

10 May '14

Dear fellow deverlopers, I’ll be attending ZuriHac in a month time, without a project of my own that I plan to pursue. Hence I’m considering to run a “GHC bug squashing” project there. Besides getting bugs squashed, this should hopefully attract new contributors and allow us to teach them the ropes and tricks and the need for notes efficiently by shoulder-surfing and face-to-face discussions. Who of you is going to attend ZuriHac as well and will want to join me in managing this project? Also, this means its a good time to start tagging tickets as suitable for beginners. I know we have a list somewhere, but googling for "ghc easy tickets" does not find it... but clicking through the wiki I find https://ghc.haskell.org/trac/ghc/wiki/Newcomers#Fixingabug which is unfortunately quite small. I’ll also look for tickets of difficulty easy and moderate. Greetings, Joachim -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de • http://www.joachim-breitner.de/ Jabber: nomeata(a)joachim-breitner.de • GPG-Key: 0xF0FBF51F Debian Developer: nomeata(a)debian.org

5 5

cgrun051
by Simon Peyton Jones 10 May '14

10 May '14

I'm getting Unexpected failures: . cgrun051 [bad exit code] (normal,hpc,optasm,profasm,ghci,threaded1,threaded2,dyn,profthreaded,optllvm,g1) =====> cgrun051(optasm) 46 of 98 [0, 2, 0] cd . && '/5playpen/simonpj/HEAD-2/inplace/bin/ghc-stage2' -fforce-recomp -dcore-lint -dcmm-lint -dno-debug-output -no-user-package-db -rtsopts -fno-ghci-history -o cgrun051 cgrun051.hs -O -fasm >cgrun051.comp.stderr 2>&1 cd . && ./cgrun051 </dev/null >cgrun051.run.stdout 2>cgrun051.run.stderr Wrong exit code (expected 0 , actual 1 ) cgrun051 throws an exception, so presumably should have exit code 1. And it does. But the test seems to test for exit code 0. Why does the test check for 0 not 1? And what has changed? This is HEAD. I don't have a virgin head right now, but I don't think any of my changes could cause this. I am also seeing the ffi/should_compile cc004 [stderr mismatch] (normal) errors that Joachim reported. It'd be great if someone could fix that, whatever it is. SImon

3 2

GHC status report
by Simon Peyton Jones 09 May '14

09 May '14

As Austin has told us, there's a draft of the GHC Status Report for the HCAR, here: https://ghc.haskell.org/trac/ghc/wiki/Status/May14 Have we missed out something you have been working hard on? Do take a moment to add a bullet in an appropriate place (it's a wiki). I'd like to be sure that we are giving credit to all the appropriate people, so please help us fix that too. GHC is a team effort. Deadline is 1 May I think. Thanks Simon

10 40

GHC not parsing TH pretty printed code
by George Roldugin 09 May '14

09 May '14

Hi all, I've hit a problem with GHC 7.8 where the code pretty printed by Template Haskell fails to parse. The problem is with the placement of explicit semicolons in place of indentation. The following code pretty printed by TH won't parse: bar = do {let x = 5; return x} GHC expects the following placement of semicolons: bar = do {let x = 5 ; return x} I reported this bug and included sample TH code to reproduce the problem here: https://ghc.haskell.org/trac/ghc/ticket/9022 I rely on compiling code constructed with TH for my thesis project. Is there a work around I could use besides going back to GHC 7.6? Cheers, George

3 2

Simplifying casts and beta-reduction
by Conal Elliott 08 May '14

08 May '14

I'm looking for tools to help simplify Core terms that involve casts. In particular, I want to beta-reduce when the function is wrapped with a cast. Do the GHC sources have such utility functions? Here is an example of a lambda expression with a cast. In context, it's applied to two type arguments and two dictionary arguments and returns a function of one more argument. (The function `(-->)` is defined as `(f --> h) g = h . g . f`.) > ((\ (@ a8) (@ b) > ($dEncodable :: Encodable a8) > ($dEncodable1 :: Encodable b) -> > case $dEncodable of _ > D:Encodable _ tpl_B3 -> > case $dEncodable1 of _ > D:Encodable tpl_B2 _ -> > \ (g :: a8 -> b) (eta :: Encode a8) -> tpl_B2 (g (tpl_B3 eta)) > ) > `cast` (forall a8 b. > <Encodable a8>_R > -> <Encodable b>_R > -> <a8 -> b>_R > -> Sub (Sym (TFCo:R:Encode(->)[0] <a8>_N <b>_N)) > :: (forall a8 b. (Encodable a8, Encodable b) => > (a8 -> b) -> Encode a8 -> Encode b) > ~# > (forall a8 b. (Encodable a8, Encodable b) => > (a8 -> b) -> Encode (a8 -> b)))) I can imagine pushing the `cast` down through the type lambdas while stripping off `forall` coercions, then pushing the remaining `cast` through the dictionary arguments, while stripping off the outer `<Encodable t>_R ->` coercion wrappers, and then pushing the cast into the `case` alternatives and the lambda body, leaving something like > (\ (@ a8) (@ b) > ($dEncodable :: Encodable a8) > ($dEncodable1 :: Encodable b) -> > case $dEncodable of _ > D:Encodable _ tpl_B3 -> > case $dEncodable1 of _ > D:Encodable tpl_B2 _ -> > \ (g :: a8 -> b) -> > (\ (eta :: Encode a8) -> tpl_B2 (g (tpl_B3 eta))) > `cast` (Sub (Sym (TFCo:R:Encode(->)[0] <a8>_N <b>_N)) > :: (Encode a8 -> Encode b) > ~# > (Encode (a8 -> b)))) Now, given type, dictionary, and function arguments, I think we could beta-reduce. Before I try implementing these transformations, does something like them already exist in GHC? Thanks, -- Conal

3 8

Cabal can't open a TCP connection
by Simon Peyton Jones 08 May '14

08 May '14

Dear ghc-devs Cabal can't open a TCP connection from my Linux box. It used to work fine. Here's what it says: cabal -v3 update Downloading the latest package list from hackage.haskell.org Sending: GET http://hackage.haskell.org/packages/archive/00-index.tar.gz HTTP/1.1 User-Agent: cabal-install/1.16.0.2 Host: hackage.haskell.org proxy uri host: , port: Creating new connection to cabal: openTCPConnection: host lookup failure for "" simonpj@cam-05-unx:~/code/HEAD$ cabal --version cabal-install version 1.16.0.2 using version 1.16.0.3 of the Cabal library simonpj@cam-05-unx:~/code/HEAD$ Can anyone help me work out what to do? git works just fine, so the internet is connected all right. Thanks Simon

4 5

SV: Cabal can't open a TCP connection
by Niklas Larsson 08 May '14

08 May '14

Looks like your proxy configuration is wrong. Git has its own proxy settings in .gitconfig. Look at the http_proxy environment variable. If you're not actually behind a proxy the problem could be the existence of http_proxy or HTTP_PROXY. ----- Ursprungligt meddelande ----- Från: "Simon Peyton Jones" <simonpj(a)microsoft.com> Skickat: ‎2014-‎05-‎08 14:04 Till: "ghc-devs(a)haskell.org" <ghc-devs(a)haskell.org> Kopia: "Simon Marlow" <marlowsd(a)gmail.com> Ämne: Cabal can't open a TCP connection Dear ghc-devs Cabal can’t open a TCP connection from my Linux box. It used to work fine. Here’s what it says: cabal -v3 update Downloading the latest package list from hackage.haskell.org Sending: GET http://hackage.haskell.org/packages/archive/00-index.tar.gz HTTP/1.1 User-Agent: cabal-install/1.16.0.2 Host: hackage.haskell.org proxy uri host: , port: Creating new connection to cabal: openTCPConnection: host lookup failure for "" simonpj@cam-05-unx:~/code/HEAD$ cabal --version cabal-install version 1.16.0.2 using version 1.16.0.3 of the Cabal library simonpj@cam-05-unx:~/code/HEAD$ Can anyone help me work out what to do? git works just fine, so the internet is connected all right. Thanks Simon

1 0

Automating Windows setup for development
by Niklas Larsson 06 May '14

06 May '14

Hi! I have made a Powershell script that sets up a GHC build environment on Windows from scratch. It's at https://github.com/melted/getghc. It largely follows the instructions at https://ghc.haskell.org/trac/ghc/wiki/Building/Preparation/Windows/MSYS2. It has no dependencies aside from Powershell 3.0. The reason I made it was to make it easier to set up a build bot on a clean machine. But it's probably generally useful for people who want to develop GHC on Windows. Usage: - Put the script in an empty directory - Run it - Hope for the best Potential problems: - If Python 2.7 is already installed and not on the path, the script wont install it, and msys wont pick up the path. Solution: Add the path to python in the control panel. - Powershell wont run unsigned scripts by default. The error message will give a hint how to disable that. I will add a signed version as soon as I can figure out how. Niklas

5 5