- ghc-devs - Haskell.org

Q: Types in GADT pattern match
by Gabor Greif 30 Oct '17

30 Oct '17

Hi Devs! I encountered a curious restriction with type signatures (tyvar bindings) in GADT pattern matches. GHC won't let me directly capture the refined type structure of GADT constructors like this: {-# Language GADTs, ScopedTypeVariables #-} data Foo a where Bar :: Foo [a] foo :: Foo a -> () foo b@(Bar :: Foo [a]) = quux b where quux :: Foo [b] -> () quux Bar = () I get: *test.hs:7:8: **error:* * • Couldn't match type ‘a1’ with ‘[a]’* * ‘a1’ is a rigid type variable bound by* * the type signature for:* * foo :: forall a1. Foo a1 -> ()* * at test.hs:6:1-18* * Expected type: Foo a1* * Actual type: Foo [a]* To me it appears that the type refinement established by the GADT pattern match is not accounted for. Of course I can write: foo :: Foo a -> () foo b@Bar | (c :: Foo [a]) <- b = quux c where quux :: Foo [b] -> () quux Bar = () but it feels like a complicated way to do it... My question is whether this is generally seen as the way to go or whether ScopedTypeVariables coming from a GADT pattern match should be able to capture the refined type. To me the latter seems more useful. Just wanted to feel the waters before writing a ticket about this. Cheers and thanks, Gabor

6 8

FW: How does GHC's testsuite work?
by Simon Peyton Jones 30 Oct '17

30 Oct '17

Forwarding to ghc-devs, a better list for this question. Simon -----Original Message----- From: Glasgow-haskell-users [mailto:glasgow-haskell-users-bounces@haskell.org] On Behalf Of Sébastien Hinderer Sent: 30 October 2017 15:18 To: glasgow-haskell-users(a)haskell.org Subject: How does GHC's testsuite work? Dear all, I am a member of OCaml's developement team. More specifically, I am working on a test-driver for the OCaml compiler, which will be part of OCaml's 4.06 release. I am currently writing an article to describe the tool and its principles. In this article, I would like to also talk about how other compilers' testsuites are handled and loking how things are done in GHC is natural. In OCaml, our testsuite essentially consist in whole programs that we compile and run, checking that the compilation and execution results match the expected ones. From what I could see from GHC's testsuite, it seemed to me that it uses Python to drive the tests. I also understood that the testsuite has tests that are more kind of unit-tests, in the .T file. Am I correct here? Or do you guys also have whole program tests? If you do, how do you compile and run them? Any comment / hint on this aspect of the test harness' design would be really helpful. Many thanks in advance, Sébastien. _______________________________________________ Glasgow-haskell-users mailing list Glasgow-haskell-users(a)haskell.org https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmail.haskel…

1 0

Family instance consistency
by Simon Peyton Jones 29 Oct '17

29 Oct '17

Edward Could you possibly take a look at #14396? It's going to bite us soon, as Alan's Trees That Grow patch lands (see the branch mentioned on the ticket). (In fact, Alan, I don't know how you build that branch ; I can't build it.) Simon

2 1

RE: [commit: ghc] master: Make tagForCon non-linear (faf60e8)
by Simon Peyton Jones 27 Oct '17

27 Oct '17

Bartosz Why not just cache the #datacons in the TyCon (in the DataTyCon constructor). That seems simple, direct, and fast. And simpler than the stuff you are forced into here. Simon | -----Original Message----- | From: ghc-commits [mailto:ghc-commits-bounces@haskell.org] On Behalf Of | git(a)git.haskell.org | Sent: 27 October 2017 23:03 | To: ghc-commits(a)haskell.org | Subject: [commit: ghc] master: Make tagForCon non-linear (faf60e8) | | Repository : ssh://git@git.haskell.org/ghc | | On branch : master | Link : | https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fghc.haske | ll.org%2Ftrac%2Fghc%2Fchangeset%2Ffaf60e858a293affca463043c830e1edb568500 | 3%2Fghc&data=02%7C01%7Csimonpj%40microsoft.com%7C677da6a867674476306508d5 | 1d869247%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636447386176303711& | sdata=tp4PTSkN5iBwKkWgZdT%2BcHKK8EKz8NQZeE%2FegVdx%2FIM%3D&reserved=0 | | >--------------------------------------------------------------- | | commit faf60e858a293affca463043c830e1edb5685003 | Author: Bartosz Nitka <bnitka(a)fb.com> | Date: Fri Oct 20 20:30:52 2017 +0100 | | Make tagForCon non-linear | | Computing the number of constructors for TyCon is linear | in the number of constructors. | That's wasteful if all you want to check is if that | number is smaller than what fits in tag bits | (usually 8 things). | | What this change does is to use a function that can | determine the ineqaulity without computing the size. | | This improves compile time on a module with a | data type that has 10k constructors. | The variance in total time is (suspiciously) high, | but going by the best of 3 the numbers are 8.186s vs 7.511s. | For 1000 constructors the difference isn't noticeable: | 0.646s vs 0.624s. | The hot spots were cgDataCon and cgEnumerationTyCon | where tagForCon is called in a loop. | | One alternative would be to pass down the size. | | Test Plan: harbormaster | | Reviewers: bgamari, simonmar, austin | | Reviewed By: simonmar | | Subscribers: rwbarton, thomie | | Differential Revision: https://phabricator.haskell.org/D4116 | | | >--------------------------------------------------------------- | | faf60e858a293affca463043c830e1edb5685003 | compiler/codeGen/StgCmmClosure.hs | 11 ++++++++--- | compiler/types/TyCon.hs | 16 +++++++++++++++- | 2 files changed, 23 insertions(+), 4 deletions(-) | | diff --git a/compiler/codeGen/StgCmmClosure.hs | b/compiler/codeGen/StgCmmClosure.hs | index 1da1f70..2501ec9 100644 | --- a/compiler/codeGen/StgCmmClosure.hs | +++ b/compiler/codeGen/StgCmmClosure.hs | @@ -361,13 +361,18 @@ type DynTag = Int -- The tag on a *pointer* | isSmallFamily :: DynFlags -> Int -> Bool isSmallFamily dflags fam_size | = fam_size <= mAX_PTR_TAG dflags | | +-- | Faster version of isSmallFamily if you haven't computed the size | yet. | +isSmallFamilyTyCon :: DynFlags -> TyCon -> Bool isSmallFamilyTyCon | +dflags tycon = | + tyConFamilySizeAtMost tycon (mAX_PTR_TAG dflags) | + | tagForCon :: DynFlags -> DataCon -> DynTag tagForCon dflags con | - | isSmallFamily dflags fam_size = con_tag | - | otherwise = 1 | + | isSmallFamilyTyCon dflags tycon = con_tag | + | otherwise = 1 | where | con_tag = dataConTag con -- NB: 1-indexed | - fam_size = tyConFamilySize (dataConTyCon con) | + tycon = dataConTyCon con | | tagForArity :: DynFlags -> RepArity -> DynTag tagForArity dflags arity | diff --git a/compiler/types/TyCon.hs b/compiler/types/TyCon.hs index | 39d2e9b..103c824 100644 | --- a/compiler/types/TyCon.hs | +++ b/compiler/types/TyCon.hs | @@ -78,7 +78,7 @@ module TyCon( | tyConDataCons, tyConDataCons_maybe, | tyConSingleDataCon_maybe, tyConSingleDataCon, | tyConSingleAlgDataCon_maybe, | - tyConFamilySize, | + tyConFamilySize, tyConFamilySizeAtMost, | tyConStupidTheta, | tyConArity, | tyConRoles, | @@ -2205,6 +2205,20 @@ tyConFamilySize tc@(AlgTyCon { algTcRhs = rhs }) | _ -> pprPanic "tyConFamilySize 1" | (ppr tc) | tyConFamilySize tc = pprPanic "tyConFamilySize 2" (ppr tc) | | +-- | Determine if number of value constructors a 'TyCon' has is smaller | +-- than n. Faster than tyConFamilySize tc <= n. | +-- Panics if the 'TyCon' is not algebraic or a tuple | +tyConFamilySizeAtMost :: TyCon -> Int -> Bool tyConFamilySizeAtMost | +tc@(AlgTyCon { algTcRhs = rhs }) n | + = case rhs of | + DataTyCon { data_cons = cons } -> lengthAtMost cons n | + NewTyCon {} -> 1 <= n | + TupleTyCon {} -> 1 <= n | + SumTyCon { data_cons = cons } -> lengthAtMost cons n | + _ -> pprPanic "tyConFamilySizeAtMost | 1" | + (ppr tc) | +tyConFamilySizeAtMost tc _ = pprPanic "tyConFamilySizeAtMost 2" (ppr | +tc) | + | -- | Extract an 'AlgTyConRhs' with information about data constructors | from an | -- algebraic or tuple 'TyCon'. Panics for any other sort of 'TyCon' | algTyConRhs :: TyCon -> AlgTyConRhs | | _______________________________________________ | ghc-commits mailing list | ghc-commits(a)haskell.org | https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmail.hask | ell.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Fghc- | commits&data=02%7C01%7Csimonpj%40microsoft.com%7C677da6a867674476306508d5 | 1d869247%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C636447386176313720& | sdata=NfTqei3XryWnCiNRgOZDT6XfMRMfAST7EKFBSi%2FCXa8%3D&reserved=0

2 2

FW: [commit: ghc] wip/nfs-locking: Use conventional whitespacing for @. (31515fa)
by Simon Peyton Jones 27 Oct '17

27 Oct '17

I've received 1500 of these commit messages overnight, and another arrives every few seconds. Are they likely to stop soon? Simon -----Original Message----- From: ghc-commits [mailto:ghc-commits-bounces@haskell.org] On Behalf Of git(a)git.haskell.org Sent: 27 October 2017 01:06 To: ghc-commits(a)haskell.org Subject: [commit: ghc] wip/nfs-locking: Use conventional whitespacing for @. (31515fa) Repository : ssh://git@git.haskell.org/ghc On branch : wip/nfs-locking Link : https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fghc.haskell… >--------------------------------------------------------------- commit 31515fad107d28f83b47d6249dd7b8c1eeb3bc70 Author: Andrey Mokhov <andrey.mokhov(a)gmail.com> Date: Fri Feb 26 11:37:47 2016 +0000 Use conventional whitespacing for @. See #210. >--------------------------------------------------------------- 31515fad107d28f83b47d6249dd7b8c1eeb3bc70 src/GHC.hs | 2 +- src/Rules/Actions.hs | 2 +- src/Rules/Compile.hs | 2 +- src/Rules/Data.hs | 4 ++-- src/Rules/Dependencies.hs | 2 +- src/Rules/Documentation.hs | 2 +- src/Rules/Generate.hs | 2 +- src/Rules/Library.hs | 4 ++-- src/Rules/Program.hs | 6 +++--- src/Rules/Register.hs | 2 +- src/Settings/Paths.hs | 10 +++++----- src/Way.hs | 2 +- 12 files changed, 20 insertions(+), 20 deletions(-) Diff suppressed because of size. To see it, use: git diff-tree --root --patch-with-stat --no-color --find-copies-harder --ignore-space-at-eol --cc 31515fad107d28f83b47d6249dd7b8c1eeb3bc70 _______________________________________________ ghc-commits mailing list ghc-commits(a)haskell.org https://na01.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmail.haskel…

3 3

Re: [commit: ghc] wip/float-join-points: Wip on floating join points (128e4c1)
by Joachim Breitner 27 Oct '17

27 Oct '17

Hi Simon, Am Freitag, den 27.10.2017, 15:24 +0000 schrieb git(a)git.haskell.org: > commit 128e4c1ffa29f3dfade7128152c143cd601aaa3a > Author: Simon Peyton Jones <simonpj(a)microsoft.com> > Date: Fri Oct 27 16:20:24 2017 +0100 > > Wip on floating join points do you expect to merge that soon, or can I land Exitification before that? (I’m worried that I have to re-do a lot of performance analysis if the Simplifier suddenly does new things to join points). Blocking for Exitification is: * https://phabricator.haskell.org/D4130 needs to land Or rather, I’d prefer if someone who understands the Simplifier better than I fixes it properly. See https://phabricator.haskell.org/D4130#inline-32243 for bits that I am puzzled about. * I need to understand better why inlining join points in the final simplifier run regresses things. Unless you permit merging Exitification without final inlining even before we completely solve this mystery, on the grounds that - the nofib results say it’s the right thing to do and - there is a plausible theory (exit join points in unfolding enable optimizations in later modules in earlier phases). Greetings, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

1 0

wip/nfsl-locking
by Simon Peyton Jones 27 Oct '17

27 Oct '17

Can anyone switch off this torrent of commit messages for wip/nfs-locking? I'm submerged in junk. I'm about to write a junk-mail rule Simon

2 1

Please add your HCAR submissions
by Ben Gamari 27 Oct '17

27 Oct '17

Hello everyone, I am currently putting together the GHC's HCAR submission for the coming publication. If you have a something that you would like mentioned (compiler feature or otherwise) please do add it here [1] sometime before Sunday, 5 November. Alternatively, feel free to mention it to me via email or IRC. Cheers, - Ben [1] https://ghc.haskell.org/trac/ghc/wiki/Status/Oct17

1 0

Simplify.hs depends on typechecker
by Joachim Breitner 26 Oct '17

26 Oct '17

Hi, while rebuilding Simplify.o only I noticed that all of the type checker, and HsSyn stuff, and renamer stuff, and so needed to be rebuilt, which I found strange. After a little investigation, it seems that the simplifier depends on CoreMonad, and that pulls some very few type-checker related things: 1. import TcRnMonad ( initTcForLookup ) import {-# SOURCE #-} TcSplice ( lookupThName_maybe ) for thNameToGhcName :: TH.Name -> CoreM (Maybe Name) thNameToGhcName th_name = do hsc_env <- getHscEnv liftIO $ initTcForLookup hsc_env (lookupThName_maybe th_name) which is not even used in GHC, but only in GHC Plugins, so this could probably be moved to a separate module pulled in by GhcPlugins.hs 2. import TcEnv ( lookupGlobal ) for instance MonadThings CoreM where lookupThing name = do { hsc_env <- getHscEnv ; liftIO $ lookupGlobal hsc_env name } This might be a bit harder to disentangle. But if successful, it would probably make building GHC in parallel quite a bit faster. And it just seems strange to me that the Core-to-Core code should depend on the type checker… I’m sending this out there in case someone has dug in that direction before and has insights to share. Greetings, Joachim (Attached is a transitively reduced dependency graph of Simplify.hs, created using Iavor’s https://github.com/yav/graphmod/wiki) -- Joachim “nomeata” Breitner mail(a)joachim-breitner.de https://www.joachim-breitner.de/

2 1

Bringing back #ghc IRC logging
by Niklas Hambüchen 25 Oct '17

25 Oct '17

Hi, as you may have noticed, #ghc IRC logging has been broken for months (though the channel topic still says "Logs: http://ircbrowse.net/ghc"). Logs being unavailable is bad because many problems and topics that were googleable before are no longer googleable. We have discussed in the last 2 days on #ghc whether we should just join #ghc with a 0-effort-on-our-side public logging service like botbot.me. This one in particular doesn't have the best UI but I found it to be very googleable, and many open-source projects use it successfully. Is anybody opposed to this? Niklas

4 5