- ghc-devs - Haskell.org

abi-depends field
by Simon Peyton Jones 22 May '18

22 May '18

I'm seeing lots of these errors. What's this "ignoring (possibly broken) abi-depends field" stuff? Or have I failed to update sufficiently vigorously somwhow? Simon =====> T13350(normal) 1 of 1 [0, 0, 0] cd "./should_compile/T13350/T13350.run" && $MAKE -s --no-print-directory T13350 Actual stdout output differs from expected: diff -uw "/dev/null" "./should_compile/T13350/T13350.run/T13350.run.stdout.normalised" --- /dev/null 2018-04-26 12:34:15.027999213 +0100 +++ ./should_compile/T13350/T13350.run/T13350.run.stdout.normalised 2018-05-21 10:33:05.181794031 +0100 @@ -0,0 +1 @@ +ignoring (possibly broken) abi-depends field for packages *** unexpected failure for T13350(normal) Unexpected results from: TEST="T13350" SUMMARY for test run started at Mon May 21 10:32:53 2018 BST 0:00:12 spent to go through 1 total tests, which gave rise to 1 test cases, of which 0 were skipped 0 had missing libraries 0 expected passes 0 expected failures 0 caused framework failures 0 caused framework warnings 0 unexpected passes 1 unexpected failures 0 unexpected stat failures Unexpected failures: should_compile/T13350/T13350.run T13350 [bad stdout] (normal)

3 3

505 Error on https://www.haskell.org/ghc/license.html
by Joachim Breitner 22 May '18

22 May '18

Hi, on https://www.haskell.org/ghc/license.html I see 500 Internal Server Error nginx before and after the text. I don’t expect many people to look there, but still ugly. Cheers, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

2 1

Potential improvements for CSE, strictness analyzer, let-floating
by Peter Podlovics 21 May '18

21 May '18

Hi all, During the summer, as a university project, I would like to make some contributions to GHC. There are three topics in particular that piqued my interest: common subexpression elimination, strictness analysis, and let-floating. I would like to ask you whether there is any room for improvement in these parts of the compiler. Could you give me some pointers? Thanks in advance, Peter

3 3

Built-in type family with one input argument
by alice 21 May '18

21 May '18

Hello, I’ve made a built-in function that takes a list and returns a list: `type family MySet [k] -> [k]`. I think I did something wrong because right now I’m getting an exception like that: ``` Extracted [Int, Bool, Char] — input list 1 List_ty '[Bool, Char, Int] — sorted list 1 Extracted [Int, Bool, Char] — input list 2 (?) List_ty '[Bool, Char, Int] — sorted list 2 (?) ghc: panic! (the 'impossible' happened) (GHC version 8.4.2 for x86_64-apple-darwin): ASSERT failed! Bad coercion hole co_a2QJ: MySet '[Bool, Char, Int] nominal (MySet '[Int, Bool, Char] :: [*]) ~# ('[Bool, Char, Int] :: [*]) Call stack: CallStack (from HasCallStack): callStackDoc, called at compiler/utils/Outputable.hs:1208:22 in ghc:Outputable assertPprPanic, called at compiler/typecheck/TcMType.hs:306:25 in ghc:TcMType Call stack: CallStack (from HasCallStack): callStackDoc, called at compiler/utils/Outputable.hs:1150:37 in ghc:Outputable pprPanic, called at compiler/utils/Outputable.hs:1206:5 in ghc:Outputable assertPprPanic, called at compiler/typecheck/TcMType.hs:306:25 in ghc:TcMType Please report this as a GHC bug: http://www.haskell.org/ghc/reportabug ``` when I’m trying to check the ~ relation: ``` test :: Proxy (GHC.TypeLits.MySet '[Int, Bool, Char]) -> Proxy (GHC.TypeLits.MySet '[Int, Bool, Char]) test = id ``` But when I do something like that: ``` test :: (GHC.TypeLits.MySet '[Int, Bool, Char] ~ as, GHC.TypeLits.MySet '[Int, Bool, Char] ~ bs) => Proxy as -> Proxy bs test = id ``` everything is fine. I think the problem is that I tried to follow TcTypeNats#typeSymbolCmpTyCon etc style but all the functions there have two inputs and one output with built-in kind (I tried to do [k] kind by myself), so probably I understood something wrong. Could someone give me a hint what I did wrong, please? I can show my code if needed. Thank you for your time, Alice.

1 0

Re: Add HeapView functionality (ec22f7d)
by Joachim Breitner 20 May '18

20 May '18

Hi, Am Sonntag, den 20.05.2018, 15:43 +0000 schrieb git(a)git.haskell.org: > commit ec22f7ddc81b40a9dbcf140e5cf44730cb776d00 > Author: Patrick Dougherty <patrick.doc(a)ameritech.net> > Date: Wed May 16 16:50:13 2018 -0400 > > Add HeapView functionality > > This pulls parts of Joachim Breitner's ghc-heap-view library inside GHC. > The bits added are the C hooks into the RTS and a basic Haskell wrapper > to these C hooks. The main reason for these to be added to GHC proper > is that the code needs to be kept in sync with the closure types > defined by the RTS. It is expected that the version of HeapView shipped > with GHC will always work with that version of GHC and that extra > functionality can be layered on top with a library like ghc-heap-view > distributed via Hackage. whoohoo, this is great! I did not not expect that the horrifying (but maybe not horrible) hack that I wrote so many years ago becomes an official feature :-) I’ll be curious in what ways our community will use (or abuse) this feature… Patrick, you mention that a library like ghc-heap-view will still be required. Are you interested in updating it to use the new interfaces, or even take it over completely? Compatibility with ghc-vis (http://felsin9.de/nnis/ghc-vis/) is probably the most important goal here. I’m CC’ing the author of ghc- vis, Dennis. Or maybe ghc-vis should use the GHC-provided wrappers directly? Dennis, are you still invested in maintaining ghc-vis? Cheers, Joachim -- Joachim Breitner mail(a)joachim-breitner.de http://www.joachim-breitner.de/

1 0

TTG hsSyn for Batch and Interactive Parsing
by Alan & Kim Zimmerman 18 May '18

18 May '18

I want to be able to run the GHC parser in one of two modes, batch which functions as now, and interactive which will (eventually) be incremental. In addition, the hsSyn AST for each will have different TTG[1] annotations, so that it can better support IDE usage. I think this can be done by changing the types in HsExtension to introduce a 'Process' type as follows data Pass = Parsed Process | Renamed | Typechecked deriving (Data) data Process = Batch | Interactive deriving (Show, Data) We then rename the pass synonyms so that batch is the default type GhcPs = GhcPass ('Parsed 'Batch) type GhcPsI = GhcPass ('Parsed 'Interactive) I have attached a simple proof of concept file, which emulates parsing and renaming. Is this an appropriate approach to take? Alan [1] Trees That Grow https://ghc.haskell.org/trac/ghc/wiki/ ImplementingTreesThatGrow

2 9

Phabricator reboot
by Ben Gamari 17 May '18

17 May '18

Hi everyone, Due to some strange behavior from our Phabricator instance over the last few days I'm going to be bringing it down for a quick reboot in about 30 minutes. The downtime should only be a couple of minutes long. Cheers, - Ben

1 0

HIE Files
by Zubin Duggal 15 May '18

15 May '18

Hello, I will be working on a GSOC project that will allow GHC to output a new .hie file to be written next to .hi files. It will contain information about the typechecked Haskell AST, allowing tooling(like haddocks --hyperlinked-source and haskell-ide-engine) to work without having to parse, rename and typecheck files all over again. I have made a GHC wiki page containing more details here: https://ghc.haskell.org/trac/ghc/wiki/HIEFiles Looking forward to any comments and suggestions. Thanks, Zubin.

4 6

How does the IO manager handle reading regular files?
by Niklas Hambüchen 14 May '18

14 May '18

I just got reminded that epoll() has no effect on regular files on Linux by reading an nginx article [1] [2] and why that is [3] [4]. By what means does the IO manager make reads (wraps around the read() syscall on Linux) non-blocking? Does it always use read() in `foreign import safe` (or `interruptible`) so that an OS thread is spawned? It would be great if somebody could point me to the code where that's done (not again: for *regular* files). Thanks! Niklas [1]: https://www.nginx.com/blog/thread-pools-boost-performance-9x/ [2]: https://stackoverflow.com/questions/8057892/epoll-on-regular-files [3]: https://jvns.ca/blog/2017/06/03/async-io-on-linux--select--poll--and-epoll/ [4]: https://groups.google.com/forum/#!topic/comp.os.linux.development.system/K-…

2 4

Motivation for refineDefaultAlt
by Matthew Pickering 11 May '18

11 May '18

Hi all, Does anyone know the motivation for refineDefaultAlt? The comment states - -- | Refine the default alternative to a 'DataAlt', if there is a unique way to do so. OK - so the code transforms something like case x of { DEFAULT -> e } ===> case x of { Foo a1 a2 a3 -> e } but why is this necessary or desirable? Perhaps you know Simon (Jakobi)? Cheers, Matt

3 3