- ghc-devs - Haskell.org

RE: Hadrian
by Simon Peyton Jones 15 Apr '19

15 Apr '19

sounds good to me! | -----Original Message----- | From: Moritz Angermann <moritz(a)lichtzwerge.com> | Sent: 15 April 2019 14:29 | To: Simon Peyton Jones <simonpj(a)microsoft.com> | Cc: Matthew Pickering <matthewtpickering(a)gmail.com>; Andrey Mokhov | <andrey.mokhov(a)newcastle.ac.uk>; ghc-devs(a)haskell.org | Subject: Re: Hadrian | | I guess we could add | | $root/ghc-stage1 (shell script) | $root/ghc-stage2 (shell script) | | As long as we understand that those are just convenience scripts. | | Why $root? Because we can have multiple build directories for different | configurations. Thus putting the scripts into the ghc folder would not | know which root you wanted. This for the default root this would be: | | _build/ghc-stage1 | _build/ghc-stage2 | | Does that sound like an intermittent solution? | | Cheers, | Moritz | | Sent from my iPhone | | > On 15 Apr 2019, at 8:47 AM, Simon Peyton Jones via ghc-devs <ghc- | devs(a)haskell.org> wrote: | > | > Ah. I hadn't even tried the stage2 compiler (in the stage1/ directory). | Yes, that works. But perhaps a script to get to it would be helpful, | otherwise invoking stage1 will look entirely different to invoking stage2. | Not a bid deal, I grant you! | > | > S | > | > | -----Original Message----- | > | From: Matthew Pickering <matthewtpickering(a)gmail.com> | > | Sent: 15 April 2019 13:40 | > | To: Simon Peyton Jones <simonpj(a)microsoft.com> | > | Cc: Andrey Mokhov <andrey.mokhov(a)newcastle.ac.uk>; | > | ghc-devs(a)haskell.org | > | Subject: Re: Hadrian | > | | > | Does the stage2 compiler which is found in `stage1/bin/ghc` not | > | work for you? I thought the issue was that the stage1 compiler | doesn't work. | > | | > | Matt | > | | > | On Mon, Apr 15, 2019 at 1:11 PM Simon Peyton Jones via ghc-devs | > | <ghc- devs(a)haskell.org> wrote: | > | > | > | > as a temporary workaround we could create a top-level wrapper | > | script, say `ghc-stage1.sh` that will call Stage1 GHC with the | > | right arguments, just like Hadrian itself does during the build. | > | > | > | > That sound fine, thanks. Same for ghc-stage2.sh? | > | > | > | > My difficulty is that as of today I simply do not know how to | > | invoke my freshly built GHC! | > | > | > | > Simon | > | > | > | > | > | > | > | > From: Andrey Mokhov <andrey.mokhov(a)newcastle.ac.uk> > Sent: 15 | > | April 2019 12:53 > To: Simon Peyton Jones <simonpj(a)microsoft.com> | > | > Cc: ghc-devs(a)haskell.org > Subject: RE: Hadrian > > > > Hi | > | Simon, > > Apologies it’s taking so long. It’s not obvious how to | > | fix this properly, and as a temporary workaround we could create a | > | top-level wrapper script, say `ghc-stage1.sh` that will call Stage1 | > | GHC with the right arguments, just like Hadrian itself does during | > | the build. | > | > | > | > Will this work for you? | > | > | > | > Cheers, | > | > Andrey | > | > | > | > | > | > | > | > From: Simon Peyton Jones [mailto:simonpj@microsoft.com] > Sent: | > | 15 April 2019 12:28 > To: Andrey Mokhov | > | <andrey.mokhov(a)newcastle.ac.uk> > Cc: ghc-devs(a)haskell.org > | > | Subject: Hadrian > > > > Andrey and other Hadrian heros > > | > | Just to say that I am 100% stalled on using Hadrian because the > | > | in-tree binary uses the wrong library files. I reported this a few | > | > weeks ago, but it still seems unchanged > > Simon > > > > | > | Bash$ ~/code/HEAD/_build/stage0/bin/ghc --version > > The Glorious | > | Glasgow Haskell Compilation System, version > 8.9.0.20190414 > > | > | simonpj@MSRC-3645512:~/tmp$ ~/code/HEAD/_build/stage0/bin/ghc -c > | > | T16566.hs > > > > T16566.hs:1:8: error: | > | > | > | > Bad interface file: | > | > /opt/ghc/8.6.4/lib/ghc-8.6.4/base-4.12.0.0/Prelude.hi | > | > | > | > mismatched interface file versions (wanted "809020190414", | got | > | > "8064") | > | > | > | > | | > | > | > | > 1 | module T16566 where | > | > | > | > | ^^^^^^ | > | > | > | > _______________________________________________ | > | > ghc-devs mailing list | > | > ghc-devs(a)haskell.org | > | > | https://nam06.safelinks.protection.outlook.com/?url=http%3A%2F%2Fmail. | > | > | > | haskell.org%2Fcgi-bin%2Fmailman%2Flistinfo%2Fghc-devs&data=02%7C | > | 01 > | > | %7Csimonpj%40microsoft.com%7Cdf9439b6d848493aca9608d6c19f79f6%7C72f9 | > | 88 > | > | bf86f141af91ab2d7cd011db47%7C1%7C0%7C636909288021992421&sdata=dd | > | sV > pnCFLP4D7dGqPtuMdv3oahBrOvRWVDlOV2GSwBU%3D&reserved=0 | > _______________________________________________ | > ghc-devs mailing list | > ghc-devs(a)haskell.org | > http://mail.haskell.org/cgi-bin/mailman/listinfo/ghc-devs

1 0

RE: Hadrian
by Andrey Mokhov 15 Apr '19

15 Apr '19

Hi Simon, Apologies it's taking so long. It's not obvious how to fix this properly, and as a temporary workaround we could create a top-level wrapper script, say `ghc-stage1.sh` that will call Stage1 GHC with the right arguments, just like Hadrian itself does during the build. Will this work for you? Cheers, Andrey From: Simon Peyton Jones [mailto:simonpj@microsoft.com] Sent: 15 April 2019 12:28 To: Andrey Mokhov <andrey.mokhov(a)newcastle.ac.uk> Cc: ghc-devs(a)haskell.org Subject: Hadrian Andrey and other Hadrian heros Just to say that I am 100% stalled on using Hadrian because the in-tree binary uses the wrong library files. I reported this a few weeks ago, but it still seems unchanged Simon Bash$ ~/code/HEAD/_build/stage0/bin/ghc --version The Glorious Glasgow Haskell Compilation System, version 8.9.0.20190414 simonpj@MSRC-3645512:~/tmp$ ~/code/HEAD/_build/stage0/bin/ghc -c T16566.hs T16566.hs:1:8: error: Bad interface file: /opt/ghc/8.6.4/lib/ghc-8.6.4/base-4.12.0.0/Prelude.hi mismatched interface file versions (wanted "809020190414", got "8064") | 1 | module T16566 where | ^^^^^^

3 3

Hadrian
by Simon Peyton Jones 15 Apr '19

15 Apr '19

Andrey and other Hadrian heros Just to say that I am 100% stalled on using Hadrian because the in-tree binary uses the wrong library files. I reported this a few weeks ago, but it still seems unchanged Simon Bash$ ~/code/HEAD/_build/stage0/bin/ghc --version The Glorious Glasgow Haskell Compilation System, version 8.9.0.20190414 simonpj@MSRC-3645512:~/tmp$ ~/code/HEAD/_build/stage0/bin/ghc -c T16566.hs T16566.hs:1:8: error: Bad interface file: /opt/ghc/8.6.4/lib/ghc-8.6.4/base-4.12.0.0/Prelude.hi mismatched interface file versions (wanted "809020190414", got "8064") | 1 | module T16566 where | ^^^^^^

1 0

Getting Rewrite Rules from ModGuts
by Bill Hallahan 12 Apr '19

12 Apr '19

Hi, I'm trying to use the GHC API (8.2.2) to extract the rewrite rules from a module, but have run into some difficulties. I've written the following code (all code is also attached as files): module Main where import GHC import GHC.Paths import HscTypes main :: IO () main = do loadProj "Test.hs" loadProj :: FilePath -> IO () loadProj src = do modgutss <- runGhc (Just libdir) $ do flags <- getSessionDynFlags _ <- setSessionDynFlags flags target <- guessTarget src Nothing _ <- setTargets [target] _ <- load LoadAllTargets mod_graph <- getModuleGraph parsed_mods <- mapM parseModule mod_graph typed_mods <- mapM typecheckModule parsed_mods desug_mods <- mapM desugarModule typed_mods return $ map coreModule desug_mods let rules = map (\mg -> ( moduleNameString . moduleName $ mg_module mg , length $ mg_rules mg) ) modgutss print $ rules return () It loads Test.hs (see below my signature) which contains a single rewrite rule. Then, it obtains the ModGuts, and prints a tuple of the name of the module, and the length of the rule list. I would expect this to be ("Test", 1), but, in fact, what gets printed is ("Test", 0). For some reason, the ModGuts does not have the rewrite rule from the file. I'm guessing I'm misunderstanding something about how the GHC API works. If anyone has any advice about how to fix this code to load rewrite rules (or even where to begin looking) I would appreciate it! Thanks, Bill Hallahan Test.hs: module Test where import Prelude hiding (map) map :: (a -> b) -> [a] -> [b] map f (x:xs) = f x:map f xs map _ [] = [] {-# NOINLINE [0] map #-} {-# RULES "map/map" forall f g xs . map f (map g xs) = map (f . g) xs #-}

3 5

[ANNOUNCE] GHC 8.6.5-rc1 is now available
by Ben Gamari 12 Apr '19

12 Apr '19

Hello everyone, The GHC team is proud to announce the first release candidate of 8.6.5. The source distribution, binary distributions, and documentation are available at https://downloads.haskell.org/~ghc/8.6.5-rc1 This release fixes a handful of issues with 8.6.4: - Binary distributions once again ship with Haddock documentation including syntax highlighted source of core libraries (#16445) - A build system issue where use of GCC with `-flto` broke `configure` was fixed (#16440) - An bug affecting Windows platforms wherein XMM register values could be mangled when entering STG has been fixed (#16514) - Several Windows packaging issues resulting from the recent CI rework. (#16408, #16398, #16516) Since this is a relatively small bugfix release we are bypassing the alpha releases and moving right to the release candidate stage. If all goes well the final release will be cut next week. As always, if anything looks amiss do let us know. Happy testing! Cheers, - Ben

4 7

Unable to build a profiled version of GHC Head
by Roland Senn 09 Apr '19

09 Apr '19

Hi, I'm unable to build a profiled version of GHC Head. I did: git clone --recursive https://gitlab.haskell.org/ghc/ghc ghcprof cd ghcprof git remote set-url --add origin https://gitlab.haskell.org/RolandSenn/ghc git remote set-url --delete origin https://gitlab.haskell.org/ghc/ghc git remote add upstream https://gitlab.haskell.org/ghc/ghc git fetch upstream git push origin master hadrian/build.sh -c --flavour=prof -j2 I get the follwoing failure: ... | Run Ghc CompileCWithGhc Stage1: utils/hp2ps/Utilities.c => _build/stage1/utils/hp2ps/build/c/Utilities.dyn_o | Run Cc FindCDependencies Stage1: libraries/ghc-prim/cbits/atomic.c => _build/stage1/libraries/ghc-prim/build/c/cbits/atomic.p_o.d | Run Ghc CompileCWithGhc Stage1: libraries/ghc-prim/cbits/atomic.c => _build/stage1/libraries/ghc-prim/build/c/cbits/atomic.p_o | Run HsCpp: compiler/prelude/primops.txt.pp => _build/stage1/compiler/build/primops.txt | Run GenPrimopCode: _build/stage1/compiler/build/primops.txt => _build/stage1/libraries/ghc-prim/build/GHC/Prim.hs libraries/ghc-prim/cbits/atomic.c: In function ‘hs_atomic_nand8’: libraries/ghc-prim/cbits/atomic.c:181:10: error: note: ‘__sync_fetch_and_nand’ changed semantics in GCC 4.4 return __sync_fetch_and_nand((volatile StgWord8 *) x, (StgWord8) val); ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ | 181 | return __sync_fetch_and_nand((volatile StgWord8 *) x, (StgWord8) val); | ^ | Run Cc FindCDependencies Stage1: libraries/ghc-prim/cbits/pext.c => _build/stage1/libraries/ghc-prim/build/c/cbits/pext.dyn_o.d | Run Cc FindCDependencies Stage1: utils/hp2ps/Main.c => _build/stage1/utils/hp2ps/build/c/Main.dyn_o.d | Run Ghc CompileCWithGhc Stage1: libraries/ghc-prim/cbits/pext.c => _build/stage1/libraries/ghc-prim/build/c/cbits/pext.dyn_o | Run GenPrimopCode: _build/stage1/compiler/build/primops.txt => _build/stage1/libraries/ghc-prim/build/GHC/PrimopWrappers.hs | Run Ghc FindHsDependencies Stage1: libraries/ghc-prim/GHC/CString.hs (and 7 more) => _build/stage1/libraries/ghc-prim/.dependencies.mk | Run Ghc CompileCWithGhc Stage1: libraries/ghc-prim/cbits/pext.c => _build/stage1/libraries/ghc-prim/build/c/cbits/pext.p_o | Remove file _build/stage1/libraries/ghc-prim/.dependencies.mk.bak Error when running Shake build system: at action, called at src/Rules.hs:68:19 in main:Rules at need, called at src/Rules.hs:85:5 in main:Rules * Depends on: _build/stage1/bin/ghc at need, called at src/Utilities.hs:71:18 in main:Utilities * Depends on: _build/stage1/libraries/ghc-prim/build/libHSghc-prim-0.6.1_p-ghc8.9.20190408.so at need, called at src/Rules/Library.hs:121:5 in main:Rules.Library * Depends on: _build/stage1/libraries/ghc-prim/build/GHC/IntWord64.p_dyn_o at error, called at src/Hadrian/Oracles/TextFile.hs:65:20 in main:Hadrian.Oracles.TextFile * Raised the exception: No dependencies found for file '_build/stage1/libraries/ghc-prim/build/GHC/IntWord64.p_dyn_o' I'm using Debian 9; x64_86; GHC 8.6.3; gcc (Debian 6.3.0-18+deb9u1) 6.3.0 20170516 Any ideas what's wrong? Many thanks and kind regards Roland

2 4

Losing (parallel) sleep over module dependencies
by Matthew Pickering 09 Apr '19

09 Apr '19

Hi all, I get worried quite often about the interaction of two facts 1. Adding a new dependency to a module is easy. 2. Adding a new dependency changes the build graph and potentially reduces the amount of parallelism which the build is capable of. Has anyone thought of any strategies to analyse the dependency graph of GHC so that a linter can be created to warn against any changes which create dependencies which affect the parallelism? hs-boot files make this issue even worse I think but I'm not sure of the precise recompilation semantics. It would be quite straight forward to index ghc as it is being built to create a dependency graph (at the identifier) level but it's unclear to me whether it is possible to do anything productive with this information. Cheers, Matt

2 1

Validation Failures on aarch64
by Travis Whitaker 09 Apr '19

09 Apr '19

Hello GHC devs, When attempting to validate a patch on aarch64, it seems there are a large number of validation failures: SUMMARY for test run started at Mon Apr 8 07:19:05 2019 UTC 0:15:35 spent to go through 6890 total tests, which gave rise to 17169 test cases, of which 10018 were skipped 41 had missing libraries 3151 expected passes 150 expected failures 11 caused framework failures 0 caused framework warnings 0 unexpected passes 3798 unexpected failures 0 unexpected stat failures The failures seem consistent on recent-ish master, specifically the neighborhood of 6113d0d4540af7853c71e9f42a41c3b0bab386fd. Is this to be expected? Thanks, Travis Whitaker

4 5

Merge Request Submission Confusion
by Ara Adkins 08 Apr '19

08 Apr '19

Hey all, I'm trying to submit a minor MR as described on the wiki [0]. I can't seem to find the mentioned 'New Merge Request' button, however. Do I need a specific set of permissions to open MRs? Any pointers would be appreciated. Best, _ara [0] https://gitlab.haskell.org/ghc/ghc/wikis/home#opening-a-merge-request

3 6

CI execution
by Sylvain Henry 08 Apr '19

08 Apr '19

Hi devs, It seems that the CI doesn't check branches in GHC forks on Gitlab anymore. Is is intentional? Is there a way to trigger a CI execution manually on a specific branch? Thanks, Sylvain

2 2