May 2015 - Glasgow-haskell-users

Discovery of source dependencies without --make
by Lars Hupel 14 Dec '15

14 Dec '15

(cross-posted from [haskell-cafe]) Hi, I've got a problem in our – admittedly complex – build process. We're running an automated grading system for student-submitted homework exercises. The compilation proceeds in stages: 1) Compile a library file, marked as trustworthy 2) Compile student submissions using Safe Haskell, able to access a whitelisted set of packages and the previously compiled library files 3) Compile the test suite 4) Link everything together into an executable The invocation is similar to this: ghc -c -outputdir "$OUT" -XTrustworthy Library.hs ghc -c -outputdir "$OUT" -i"$OUT" -XSafe "$SUBMISSION" ghc -c -outputdir "$OUT" -i"$OUT" Test_Suite.hs ghc -outputdir "$OUT" -i"$OUT" -o "$OUT/runner" The second stage works well under the assumption that there's just one single submitted file. If there's more than one, I need to specify them in topological order wrt their module dependencies. Consider two simple modules: A.hs > module A where > > import B > import Library B.hs > module B where Invoking GHC with "A.hs" "B.hs" will fail: A.hs:3:1: Failed to load interface for ‘B’ My naive attempt at solving that problem was to just insert "--make" into the compiler flags. No luck, because according to [1, §4.7.3], "--make" will only look at source files and will ignore the "Library.hi" file from the previous compilation stage. If I give the path to "Library.hs" as well, GHC insists on recompiling it with changed flags ("-XSafe" instead of "-XTrustworthy"). Reading a bit further, I discovered the '-M' flag, however, not only does it output Makefile-formatted output, it also ignores .hi files. Is there any way to get the dependency discovery of '--make' without the rest? I know I could probably make it work if I made a package out of the first build stage, but I really want to avoid that in order to not increase the build complexity even further. Cheers Lars [1] <https://downloads.haskell.org/~ghc/7.6.3/docs/html/users_guide/separate-com…>

8 15

RFC: "Native -XCPP" Proposal
by Herbert Valerio Riedel 19 Jun '15

19 Jun '15

Hello *, As you may be aware, GHC's `{-# LANGUAGE CPP #-}` language extension currently relies on the system's C-compiler bundled `cpp` program to provide a "traditional mode" c-preprocessor. This has caused several problems in the past, since parsing Haskell code with a preprocessor mode designed for use with C's tokenizer has caused already quite some problems[1] in the past. I'd like to see GHC 7.12 adopt an implemntation of `-XCPP` that does not rely on the shaky system-`cpp` foundation. To this end I've created a wiki page https://ghc.haskell.org/trac/ghc/wiki/Proposal/NativeCpp to describe the actual problems in more detail, and a couple of possible ways forward. Ideally, we'd simply integrate `cpphs` into GHC (i.e. "plan 2"). However, due to `cpp`s non-BSD3 license this should be discussed and debated since affects the overall-license of the GHC code-base, which may or may not be a problem to GHC's user-base (and that's what I hope this discussion will help to find out). So please go ahead and read the Wiki page... and then speak your mind! Thanks, HVR [1]: ...does anybody remember the issues Haskell packages (& GHC) encountered when Apple switched to the Clang tool-chain, thereby causing code using `-XCPP` to suddenly break due to subtly different `cpp`-semantics?

11 14

overlapping instances in 7.10.1
by Sergei Meshveliani 16 Jun '15

16 Jun '15

Dear GHC developers, This request overrides my previous one of "7.10.1-err..." (it is simpler and more precise). The archive http://www.botik.ru/pub/local/Mechveliani/ghcQuest/7.10.1-errReport-may23-2… presents a question about ghc-7.10.1. Make it, please, with ghc-7.10.1 by ghc $doconCpOpt -O --make Main , $doconCpOpt = -fwarn-unused-matches -fwarn-unused-binds -fwarn-unused-imports -fno-warn-overlapping-patterns -XRecordWildCards -XNamedFieldPuns -XFlexibleContexts -XMultiParamTypeClasses -XUndecidableInstances -XTypeSynonymInstances -XFlexibleInstances -fcontext-stack=30 as it is written there in README.txt. README.txt explains which two instances are wrongly resolved -- as I expect. In ghc-7.8.2 they are resolved in a correct way (and there is a different pragma syntax). I conclude this from running the test in docon-2.12. Am I missing something? Please, advise, ------ Sergei

2 6

Consecutive FFI calls
by David Turner 29 May '15

29 May '15

(cross-posted from ghc-dev and the cafe, with apologies to cross-subscribers) Hi, If I make a sequence of FFI calls (on a single Haskell thread) but which end up being called from different OS threads, is there any kind of ordering guarantee given? More specifically, is there a full memory barrier at the point where a Haskell thread migrates to a new OS thread? Many thanks, David

1 0

SRC_HC_OPTS in perf build
by Jeremy 27 May '15

27 May '15

build.mk.sample contains the lines: # perf matches the default settings, repeated here for comparison: SRC_HC_OPTS = -O -H64m However, in config.mk.in this is: SRC_HC_OPTS += -H32m -O What actually is the default for SRC_HC_OPTS? Why does config.mk.in seem to set it to -H32m, then every profile in build.mk.sample sets -H64m? -- View this message in context: http://haskell.1045720.n5.nabble.com/SRC-HC-OPTS-in-perf-build-tp5809863.ht… Sent from the Haskell - Glasgow-haskell-users mailing list archive at Nabble.com.

2 2

error in 7.10.1 ?
by Sergei Meshveliani 23 May '15

23 May '15

Dear GHC developers, The confirmation token letter from the bug tracker travels too long, I do not know, why. Meanwhile I cannot report there. So, I report it in the below way: http://www.botik.ru/pub/local/Mechveliani/ghcQuest/7.10.1-errReport-may22-2… Make it, please, with ghc-7.10.1 as it is written there in install.txt, including making demotest/Main. Then ./Main breaks in the middle reporting a DoCon application error ---------------------------------------------------------------------------- Factorization in K1[x], K1 a generic finite extension of a prime field K = Z/(p) .. factor f, f = (x^10 + x^7 + x^5 + x^3 + x^2 + (2)*x+ (2)) <- K[x] = ((Z/<5>)["t"]/<(t^3 + 4*t + 3)>)["x"] Domain term of K must contain a Finite field description. --------------------------------------------------------------------------- I obtain this for ghc-7.10.1 made from source by ghc-7.8.2 on Debian Linux, x-86, 64 bit. The story is as follows. docon-2.12 http://www.botik.ru/pub/local/Mechveliani/docon/2.12/ has been tested under ghc-7.8.2, and it works. Now I am trying to port it to ghc-7.10.1. This needs inserting certain pragmas for OVERLAPPING/OVERLAPPABLE into some instance declarations. So I call this port docon-2.12.1. 7.10.1-errReport-may22-2015.zip is the error report of docon-2.12.1 to ghc-7.10.1. This application has the OVERLAPPABLE pragma set to one instance and the OVERLAPPING pragma set to many instances. GHC itself warns of overlaps. Then I 1) set the pragma to the reported instance (pair), 2) recompile the corresponding module 3) rerun make configure; make build. This process has finished in `making' the application. Then, > cd demotest > ghc $doconCpOpt -O --make -rtsopts Main > ./Main produces the above error instead of the test success report. I suspect that the reason is in treating the instance overlaps (but there may be another source). Comparing docon-2.12 + ghc-7.8.2 to 7.10.1-errReport-may22-201 + ghc-7.10.1, can you, please, find: how to fix the latter version? Is this a bug in ghc-7.10.1 ? Regards, -------- Sergei

1 1

-Wall and the fail method
by Nikita Karetnikov 23 May '15

23 May '15

Can -Wall be extended to report pattern match failures in do expressions, like it does for case expressions? Prelude> :set -Wall Prelude> let f = do Just x <- return Nothing; return x Prelude> let g = case Nothing of Just x -> x <interactive>:9:9: Warning: Pattern match(es) are non-exhaustive In a case alternative: Patterns not matched: Nothing One can argue that it's similar to undefined, error, and various unsafeSomething functions, which I think should be reported as well, if possible. But these things can be found already with a simple grep while a pattern match cannot. I bet it has been discussed already, but "fail" is a terrible search term, so I cannot find anything relevant in the archives nor in the bug tracker.

2 1

Recursive loop in context simplification with GHC 7.10.1
by Joe Hendrix 22 May '15

22 May '15

Hi, I’m trying to migrate some code that compiles under GHC 7.8.4 to GHC 7.10.1. It uses the comparison operator (<=) in TypeLits, and seems to get in a recursive loop when simplifying contexts. I’ve included part of the output from "-ddump-tc-trace” once it gets stuck in the loop below. I’d appreciate any help on understanding what is happening. If it’s useful I could try to extract a minimal example showing the error, but I thought I’d try to understand what’s happening first. Thanks, Joe === flatten/flat-cache hit GHC.TypeLits.<=? [1, n_abP7[sk]] s_aifJ[fuv:0] cobox_aifK Following filled tyvar s_aifJ[fuv:0] = 1 GHC.TypeLits.<=? n_abP7[sk] flattenTyVar1 n_abP7[sk] GHC.TypeLits.Nat flatten/flat-cache hit GHC.TypeLits.<=? [1, n_abP7[sk]] s_aifJ[fuv:0] cobox_aifK Following filled tyvar s_aifJ[fuv:0] = 1 GHC.TypeLits.<=? n_abP7[sk] flattenTyVar1 n_abP7[sk] GHC.TypeLits.Nat flatten/flat-cache hit GHC.TypeLits.<=? [1, n_abP7[sk]] s_aifJ[fuv:0] cobox_aifK

1 0

documentation links broken in 7.10.1
by Evan Laforge 21 May '15

21 May '15

It seems like haddock's index.html generation is broken in 7.10.1. Namely, it creates links to e.g. Codec-Binary-UTF8-Generic.html when it should be to utf8-string-1/html/Codec-Binary-UTF8-Generic.html This is haddock 2.16.0, which comes with the binary distribution for 7.10.1. I looked at https://github.com/haskell/haddock/issues but didn't see anything related to this, which makes me think it's just me. You'd think someone would notice if all documentation links were broken. Does anyone else see this?

2 2

overlapping instances 7.10.1
by Sergei Meshveliani 21 May '15

21 May '15

People, I wrote recently about finding places to set {-# OVERLAPPING #-} when porting an application from 7.8.2 to 7.10.1. I am doing this for porting docon-2.12 from 7.8.2 to 7.10.1. And ghc-7.10.1 indeed helped me to find several places to set this pragma (instead of using the total key -XOverlappingInstances). Finally, it has come to this module: --------------------------------------------- Preprocessing library docon-2.12.1... [48 of 86] Compiling ResRing__ ( ResRing__.hs, dist/build/ResRing__.o ) ResRing__.hs:183:31: Overlapping instances for Eq (Maybe PropValue) arising from a use of ‘/=’ Matching instances: instance Eq a => Eq (Maybe a) -- Defined in ‘GHC.Base’ instance (Residue r, Eq a) => Eq (r a) -- Defined in ‘ResEuc0_’ In the expression: lookup IsGxBasis ps /= Just Yes ... ---------------------------------------------- As before, I set instance {-# OVERLAPPING #-} (Residue r, Eq a) => Eq (r a) where ... in ResEuc0_.hs. But this does not help, the compiler continues with the same report. I see the difference to previous porting process in that one of the overlapping instances is of the GHC library, so that I cannot set the needed overlap pragma for it. On the other hand, ghc-7.8.2 compiled docon-2.12 successfully. What my be a way out? Thanks, ------ Sergei

2 1