- ghc-devs - Haskell.org

Max identifier length
by Christopher Done 15 Jan '19

15 Jan '19

Hey, Quick question: Does GHC have an upper limit on the length of binding names? Cheers!

3 2

RE: GHC | Some refactoring in tcInferApps (!116)
by Simon Peyton Jones 14 Jan '19

14 Jan '19

Devs I’m at a loss for how to review GitLab changes. Richard sent me the message below. So I follow the link to “View on GitLab”, or I manually edit the URL to plain https://gitlab.haskell.org/ghc/ghc/merge_requests/116 Either way, I can’t see any comments whatsoever! It says 5/5 discussions resolved, but I can’t actually see them. There is not “toggle discussion” button anywhere. What should I do? This is quite a big problem. I suppose we could issue guidance NEVER to resolve a discussion, but that seems like the wrong conclusion. Simon From: Richard Eisenberg <gitlab(a)gitlab.haskell.org> Sent: 14 January 2019 03:03 To: Simon Peyton Jones <simonpj(a)microsoft.com> Subject: Re: GHC | Some refactoring in tcInferApps (!116) Richard Eisenberg<https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitlab.ha…> commented on a discussion on compiler/typecheck/TcHsType.hs<https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitlab.ha…>: 1022 + (vcat [ ppr ki_binder 1023 + , ppr ki_arg 1024 + , ppr (tyBinderType ki_binder) 1025 + , ppr subst 1026 + , ppr (isInvisibleBinder ki_binder) ]) 1027 + ; ty_app_err ki_arg $ nakedSubstTy subst $ 1028 + mkPiTys all_ki_binders inner_ki } 1029 + 1030 + -- no binder; try applying the substitution, or fail if that's not possible 1031 + | otherwise 1032 + = try_again_after_substing_or n subst fun inner_ki all_args $ \ substed_inner_ki -> 1033 + ty_app_err ki_arg substed_inner_ki 1034 + 1035 + go n subst fun all_ki_binders inner_ki all_args@(HsValArg arg : args) 1036 + | ki_binder : ki_binders <- all_ki_binders 1037 + = if isInvisibleBinder ki_binder See if you like how I've done this now. — View it on GitLab<https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitlab.ha…>. You're receiving this email because of your account on gitlab.haskell.org. If you'd like to receive fewer emails, you can unsubscribe<https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgitlab.ha…> from this thread or adjust your notification settings.

4 8

Reviews
by Simon Peyton Jones 12 Jan '19

12 Jan '19

When reviewing a MR, I sometimes want to look at (and even add comments to) code that isn't displayed in the MR window. With Phab there was a "show 20 more lines" link, and even "show the whole file". How do I do that in GitLab? When submitting a review, I often want to add an overall comment, not related to a particular line of code? How do I do that? Thanks Simoin

5 14

More GitLab questions
by Simon Peyton Jones 11 Jan '19

11 Jan '19

* When looking at a comment on a MR, I see some snippet of code. How can I see what module that snippet is in? Apart from scrolling up to find the file header - being careful not to miss it and scroll into the previous file. * When looking a comment, there's a little button to skip to the next unresolved comment. Good. But if I respond to the comment (without resolving it), that little button disappears. How can I find the next unresolved comment? Simon

2 3

Changes to UserSettings.hs
by Matthew Pickering 10 Jan '19

10 Jan '19

Hi all, Last night my patch got merged which changed some aspects of the hadrian settings file. If you have a custom settings file (ie hadrian/UserSettings.hs), you need to remove reference to the `stage1Only` variable and add a new variable called `finalStage` which should default to `Stage2`. https://gitlab.haskell.org/ghc/ghc/commit/6486c6e49c53e75f37ed732b38c5be7ae… Cheers, Matt

1 0

Gitlab notes
by Simon Peyton Jones 09 Jan '19

09 Jan '19

Friends I've added a Wiki page to summarise things I've learned about Gitlab. Please do correct any errors I've made! https://ghc.haskell.org/trac/ghc/wiki/WorkingConventions/GitLabSPJ Thanks Simon

5 6

Hadrian
by Simon Peyton Jones 09 Jan '19

09 Jan '19

Alp You've written a helpful wiki page describing how to use Hadrian https://gitlab.haskell.org/ghc/ghc/blob/master/hadrian/doc/make.md Could you document where build-products live? I often look at .hi files. Could you add * How to do the equivalent of "sh validate" * When validating, how to do things like this: GhcStage1HcOpts += -DDEBUG -O GhcStage2HcOpts += -Wwarn -ticky GhcLibHcOpts += -ticky -Wwarn BUILD_PROF_LIBS = NO SplitObjs = YES HADDOCK_DOCS = NO BUILD_SPHINX_HTML = NO BUILD_SPHINX_PDF = NO * After modifying the source code, how to make stage1 compiler (only). In 'make' it is $ cd compiler; make 1 * After modifying the source code, build the stage2 compiler (only). In 'make' it is $ cd compiler; make 2 Perhaps you say build -freeze1? * In the testsuite, how to use the stage1 compiler $ make stage=1 TEST=T2543 * In nofib, how to build one benchmark, with extra flags, perhaps even with the stage1 compiler $ make stage=1 EXTRA_HC_OPTS="-ticky -ddump-simpl -ddump-stg" Thanks! Simon

2 4

Blocking a task indefinitely in the RTS
by Phyx 09 Jan '19

09 Jan '19

Hi All, I'm looking for a way to block a task indefinitely until it is woken up by an external event in both the threaded and non-threaded RTS and returns a value that was stored/passed. MVar works great for the threaded RTS, but for the non-threaded there's a bunch of deadlock detection in the scheduler that would forcibly free the lock and resume the task with an opaque exception. This means that MVar and anything derived from it is not usable. STMs are more expensive but also have the same deadlock code. So again no go. The reason it looks like a deadlock to the RTS is that the "Wake-up" call in the non-threaded rts will come from C code running inside the RTS. The RTS essentially just sees all tasks blocked on it's main capability and (usually rightly so) assumes a deadlock occurred. You have other states like BlockedOnForeign etc but those are not usable as a primitive. Previous iterations of I/O managers have each invented primitives for this such as asyncRead#, but they are not general and can't be re-used, and requires a different solution for threaded and non-threaded. I have started making a new primitive IOPort for this, based on the MVar code, but this is not trivial... (currently I'm getting a segfault *somewhere* in the primitive's cmm code). The reason is that the semantics are decidedly different from what MVars guarantee. I should also mention that this is meant to be internal to base (i.e no exported). So before I continue down this path and some painful debugging..., does anyone know of a way to block a task, unblock it later and pass a value back? It does not need to support anything complicated such as multiple take/put requests etc. Cheers, Tamar

4 12

[Q] Inlining done: evtRead
by Gabor Greif 08 Jan '19

08 Jan '19

Hi Simon, a simplifier question... Roughly a year ago I started learning about imported Ids, their unfoldings etc. I have very small example program that compiles on Linux. ```haskell import GHC.Event main = print evtRead ``` `evtRead` is a newtype-wrapped Int. When you compile above program with HEAD GHC without optimisation, you'll see that `evtRead` gets passed directly to `show`. But with -O1 it's unfolding will be inlined, floated to toplevel, and dumped as static global data into the using Main module. This was not the case in GHC 8.4. Not sure about 8.6 (will check). Anyway here is the inlining notice that the simplifier gave me (-ddump-inlinings -dverbose-core2core) > Inlining done: GHC.Event.Internal.evtRead > Inlined fn: (GHC.Types.I# 1#) > `cast` (Sym (GHC.Event.Internal.N:Event[0]) > :: GHC.Types.Coercible GHC.Types.Int > GHC.Event.Internal.Event) > Cont: Stop[BoringCtxt] GHC.Event.Internal.Event > I believe this is a regression, as copies of global data can pop up in potentially many different modules. What do you think? Which change could have caused this? Cheers, Gabor

3 7

Better DWARF info for Cmm procedures?
by Ömer Sinan Ağacan 07 Jan '19

07 Jan '19

Hi, Currently debugging Cmm is a bit painful because we don't have enough debug information to map assembly to Cmm lines, so I have do the mapping manually. However I realized that when building .cmm files we actually generates some debug information, in form of "ticks": //tick src<rts/Apply.cmm:631:9-37> _c2e::I64 = I64[R1 + 32]; Here the tick says that this assignment is for this Cmm line in Apply.cmm: Words = StgAP_STACK_size(ap); I was wondering what needs to be done to generate DWARF information from those so that gdb can show Cmm line we're executing, and gdb commands like `next`, `break` etc. work. I also realize that we don't consistently generate these ticks for all Cmm lines, for example, in the same Cmm dump there isn't a tick before this line: (_c2j::I64) = call MO_Cmpxchg W64(R1, stg_AP_STACK_info, stg_WHITEHOLE_info); It's actually for Apply.cmm:646. So there are two problems: - Generate ticks for _all_ Cmm lines - Generate DWARF information from those so that gdb can show current Cmm line, and commands like `next` and `break` work. Anyone know how hard would this be to implement? I'm wondering if we could turn this into a SoC project. It's a very well defined task, and given that we have some DWARF support already, perhaps it's not too hard for a SoC. Ömer

2 3